In short: Instagram ads work for a small business when three things are already true. You sell something people want on sight, the page they land on already converts the visitors it gets, and your margin can absorb the cost of buying a customer. If any of those is missing, ads will not fix it. They will make it more expensive, faster. The honest test is not "do Instagram ads work" (they demonstrably do, for some businesses) but "does the maths work for mine". You can answer that in ten minutes with numbers you already have.
Everyone answering this question sells advertising, which is why half the internet appends "reddit" to it. Here is the version with the uncomfortable parts left in.
The three numbers that decide it
Work these out first. They decide the answer more than any targeting setting ever will.
What you can afford to pay for a customer
Gone before you see it
Materials, your time, packaging, postage, the platform’s cut
What you keep
Your ceiling. Pay more than this to win a customer and the campaign loses money, however good the engagement looks.
The one thing that lifts the ceiling: a customer who comes back. If they buy three times a year, you can afford far more to meet them the first time.
The third number is your website. If a hundred people arrive from Instagram organically and none of them buy, sending a thousand paid ones produces the same result at a cost.
If those three leave you room, ads are worth testing. If they do not, they are not your next best dollar.
When Instagram ads genuinely work
- You sell something visual. Ceramics, prints, jewellery, clothing, furniture, food. The feed rewards work that stops a thumb.
- Nobody is searching for what you make. This is the real argument for Instagram over Google. Nobody types "handmade stoneware" into a search bar, so there is no existing demand to capture. It has to be created, and that is what Meta is for.
- You have a moment. A collection, a season, a sale, a market stall next weekend. A deadline gives people a reason to act rather than save your post and forget.
- Your shop already works. People can find the size, understand the shipping, and check out on a phone without thinking.
When they do not, and this is common
- Your margin is thin. Handmade work at a low price with real materials cost is the hardest case for paid advertising.
- You sell a handful of expensive commissions a year. If you need six clients annually and each one is a conversation, referrals and the right rooms will beat ads comfortably. Ads are a volume instrument.
- You have no photographs you are proud of. Bad creative is the most common cause of a failed campaign, and it cannot be bought around. If the images are not there yet, that is the job to do first, and it is a cheaper job.
- Your product page loses people. A slow page, a surprise postage cost at checkout, photos that do not answer the obvious questions. Ads simply buy you a bigger audience for the same disappointment.
The three ways small businesses lose money
They fix the ad when the problem is after the click. Most failed campaigns are competent ads pointing at a page that does not close. The spend gets blamed on Meta, the targeting gets rebuilt, and the leak stays where it was. Look at what happens after the click first. It is free to check.
They stop too early. Meta's delivery needs an uninterrupted stretch of data before its results mean anything. Turning campaigns off after a quiet weekend, changing the budget daily, restarting with a new audience: each of those resets the clock. The hardest part of running ads is not touching them.
They test one idea in four costumes. The same message with a different button colour is not a test. Genuinely different angles tell you something. Variations of one angle tell you nothing, slowly.
What working actually looks like
Not likes, not reach, not a dashboard full of green. Working means that for every dollar in, more than a dollar came back, and you can see which idea did it. That is the whole scoreboard. If you cannot get a straight answer to "what did a sale cost me", the measurement is not set up properly.
A test that comes back negative is still a result. It tells you your price, your margin or your product page needs work before paid traffic is worth buying, and that is a cheap answer to a question that would otherwise cost you a year.
Before you spend a dollar
- Work out what you keep on a sale, and what a customer is worth over a year.
- Check that your product page converts the traffic you already have.
- Gather the photos and video you are proud of. If there are none, start there.
- Set up the measurement, and confirm it is recording real purchases.
- Decide what you can spend for several weeks without flinching, then leave it alone.
If steps one to three make you uncomfortable, that discomfort is the useful part. It is cheaper to find now than after a month of spend.
If the maths looks like it might work, the next two questions are which kind of ad to run first and what it will cost you in Australia.
Or if you would rather someone worked it out with you, that is what I do on Instagram and Facebook ads for Melbourne makers and shops.