Fifty dollars is the amount people spend on ads when they have decided to be serious but have not decided to be reckless. It is small enough to lose without a conversation at home, and large enough to feel like a proper test.
It buys three separate things, and only two of them are real.
The arithmetic, roughly
An ad platform sells you impressions in an auction. What you pay per thousand impressions depends on who you are trying to reach, where they live, what time of year it is, and how many other people want the same attention. A cheap country and a broad audience might cost a few dollars per thousand. A narrow audience of business buyers in a rich country in December can be twenty or forty times that.
So the first honest answer is that $50 buys somewhere between about two thousand and about forty thousand impressions, and I cannot tell you which without knowing your account.
Then people have to click. A click-through rate of one percent is normal on feed placements, half a percent is common, three percent means the creative is working. Put those together and $50 tends to produce somewhere between twenty and a hundred and fifty clicks. Round numbers for the sake of argument: call it sixty to eighty clicks at around seventy cents each.
Then those clicks have to become something. If your page converts at two percent, sixty-seven clicks gives you an expected 1.3 sales. Expected is doing a lot of work in that sentence. With a real two percent conversion rate and sixty-seven clicks, the chance of getting exactly zero sales is about twenty-six percent. At one percent, it is about half.
That is the part worth sitting with. A perfectly functional campaign, with a real product and a real page, returns nothing at all a quarter to half of the time at this budget. Not because you did it wrong. Because sixty-seven is a small number and coins land on their edge more often than people expect.
Also, the $50 is not $50. If you are in the EU or the UK and you have not given the platform a VAT number, VAT gets added to the invoice. In Poland that is twenty-three percent, so your fifty becomes sixty-one and a half. Currency conversion takes another bite if your card is not in the billing currency. Budget accordingly, or supply the number if you have one.
The part that buys nothing
Here is where small budgets stop behaving like big ones.
Ad platforms optimise by watching conversions and finding more people like the people who converted. Meta has said for years that an ad set needs roughly fifty conversions in a week before its delivery settles down. That is not a marketing figure, it is a statistics figure, and it applies to anything that learns from examples. At $50 total, you are not producing fifty conversions in a week. You are producing one, or none.
So the optimisation you are paying for does not happen. The system is still guessing when your money runs out. Whatever it learned about your buyers it learned from a sample of one, and it will forget it if you pause and restart. This is the single biggest gap between what a small budget looks like it is doing and what it is doing.
Retargeting is the same problem in a different coat. You cannot retarget sixty-seven visitors. The audience is too small to serve, and even where a platform will accept it, the pool refreshes so slowly that the ad shows the same person the same thing until they resent you. Retargeting starts being interesting somewhere in the thousands of visitors, and $50 does not get you within sight of that.
Split testing also does not survive at this scale. If ad A gets four sales and ad B gets one, that difference is noise. You would need hundreds of conversions per variant to know anything, and you have five in total. People run three creatives on ten dollars each, declare a winner, and scale the loser. I have no way to prove any given case, but the maths says it happens roughly as often as chance.
And $50 buys no answer to the question people actually want answered, which is whether the thing is viable. Viability is a question about repeat purchase, refund rate, and what happens at ten times the spend. One test cannot see any of that.
What it does buy
A read on the hook, mainly.
Click-through rate stabilises much faster than conversion rate, because clicks are common and sales are rare. A few thousand impressions is enough to tell you whether the first three seconds and the first line stop anybody. If two thousand people see your ad and eight click, the creative is not the thing to fix later, it is the thing to fix now. That is a useful answer and it costs about ten dollars, not fifty.
It buys a sanity check on the plumbing. Does the link work on iOS. Does the pixel fire. Does the checkout accept a card from another country. Does the confirmation email arrive. Every one of those has broken silently for someone this week, and paid traffic is a fast way to find out, because paid traffic arrives all at once and behaves like strangers rather than like friends being polite.
It buys a small number of real strangers looking at your offer, which is different in kind from your followers looking at it. Followers are kind. Strangers are not kind, and their indifference is information.
And it buys permission to stop. If a hundred strangers see the offer and none of them want it, that is not a verdict on you, but it is a data point, and it cost less than a night out. Most people spend that money instead on another month of thinking about it.
One more thing it can buy, if you set it up: email addresses. A person who gives you an email is a person you can talk to again for free. At $50, collecting twenty emails is often a better outcome than collecting one sale, because the sale ends and the list does not.
Where the click lands, which is the bit nobody budgets for
You can spend all $50 on getting someone to tap, and then lose them in the two seconds after the tap. This is the cheapest thing to fix and the last thing anyone looks at.
A link page is one honest answer to it. Relay is ours: your photo, your bio, unlimited links, headings and dividers so the page reads in an order rather than as a wall, social icons, five free themes. The free plan needs no card and does not expire. The part that matters for a paid test is the tap and view counts, because they split one number into two. If a hundred people viewed the page and eleven tapped the buy link, the ad worked and the page did not. If ninety tapped and nobody bought, the page worked and the checkout or the price did not. Without that split you are looking at a single zero and guessing which half of the funnel produced it.
Relay is also part of the problem, and it would be silly to pretend otherwise. Every page between the ad and the payment is a place people leave. If you sell one thing, the highest-converting destination for paid traffic is the checkout for that thing, not a page listing seven options including that thing. A link page earns its place when the traffic is organic and mixed, when people arrive curious rather than committed, or when you have several things and no idea which one they came for. When the ad says buy this candle and the page says here are my links, you have paid for a click and spent it on a menu. Relay has no payments and no custom domains, and we say so on the comparison page, which also lists four things Linktree does better. Pro is $30 a year or $4 a month plus VAT if you want all forty-seven themes, all eight fonts, custom colours, a background image, per-link images, scheduled links, inline players, email capture and the badge gone. Data sits in the European Union, fonts are self-hosted, there is no advertising and no third-party analytics, so nobody is quietly rebuilding a profile of the people you paid to send there.
How I would spend the fifty
One product. One audience. One or two creatives, not six. Ten dollars a day for five days, or five a day for ten, rather than fifty in an afternoon, because platforms deliver badly in the first hours and you want the campaign to have at least seen a weekday and a weekend. Note that daily minimums differ by platform and change without ceremony, and TikTok's sit higher than Meta's, so read the number in your own ads manager rather than trusting mine.
Send the click to the shortest path to payment that exists. Watch click-through rate for the creative, and watch what happens after the click for everything else. Write down what you expected before you start, because otherwise you will remember having expected whatever happened.
And hold the second fifty. The most useful thing a first test produces is usually a better ad, and the money to run it.
If $50 returns nothing, that is inside the normal range and not a signal to spend $500 proving it wrong. It is a signal to look at the offer, the price, and the page. There is a longer piece here on how long the first sale actually takes, and one on making it happen without ads, which is how most first sales happen anyway. If you have not decided where the money should change hands yet, the TikTok Shop versus Shopify versus just a link comparison is the thing to read before the ad, not after.
If you want a destination that reports back rather than swallowing the traffic in silence, Relay is free to start and takes about ten minutes. Then spend the fifty on finding out whether anyone wants the thing.
