Practical guides on diagnosing ad and store performance — written from the same methodology KPIHelm uses.
There is no single good ROAS for an industry: the number your store needs is set by your own margin first, and an industry figure only tells you where you stand against your sector's published range once both numbers measure the same thing. Work out your break-even, check what kind of ROAS the benchmark you found actually measures, and then read your category's range as a reference point, not a target.
Read the articleUsually because the video holds attention without giving a reason to act: the product, the offer or the next step comes too late, is unclear, or is never asked for — or the campaign was set up to find people who watch, not people who click. Before you remake the ad, check that "watched" and "clicked" in your report mean what you think, check what you asked Meta or TikTok to optimise for, and find the second at which viewers leave.
Read the articleUsually because Meta finds the easiest buyers first: in the first days it shows the ad to the people most ready to buy, and once they have bought or scrolled past, it keeps paying to reach the same people again. Before you change anything, check that sales really stopped in your store and not just in Meta's report, look at frequency and cost per result together, rule out a change you made yourself, and count whether three good days were ever a trend.
Read the articleUsually because the clicks are not the visits you think they are, or because the people behind them were never in a buying mood — TikTok sells attention cheaply, and a cheap click from someone mid-scroll is worth less than it looks. Check, in this order, that the clicks really reach your store, that the campaign is told to find buyers, that there are enough clicks to judge, and that the page continues what the video promised.
Read the article$20 a day is enough to find out whether an ad can sell — if all of it goes into one campaign and one sale costs you a few days of budget, not a few weeks. Decide when to stop in money, not in days: work out before you start what a customer is allowed to cost you, and judge once the ad has spent two to three times that amount.
Read the articleBecause the two count different things at different moments: Meta counts a tap on your ad, Shopify counts a visit only once its own code has run on a page that actually loaded — and between those two moments people leave, pages stall, and browsers or consent banners stop the counting. Some gap is normal; what matters is whether yours is stable or suddenly grew, and whether your sales moved with it.
Read the articleUsually one of two very different things: people really are leaving between cart and checkout — most often because of a cost or a step they did not expect — or they are not leaving at all, and your checkout event simply is not reaching the ad platform. Check the tracking first, because it decides which of the two problems you are about to fix.
Read the articleNo more than the margin that customer brings you — the first order plus what they buy again — within a window you can finance. Measure it per channel, and count new customers from your store, not from the ad platforms.
Read the articleBecause they count different things: an ad platform reports every sale it believes its ads influenced, under its own attribution rules, while Shopify records each order once. Use Shopify's numbers for money decisions and the platform's for optimizing ads — and watch whether the gap between them stays stable, because a sudden change usually means tracking changed, not demand.
Read the articleBecause a blended ROAS is an average: a strong channel can lift it above break-even while another channel loses money underneath. Spend 6,000 on Meta for 24,000 in sales (4.0x) and 4,000 on TikTok for 4,800 (1.2x), and the blend is 2.88x — above a 2.5x break-even while TikTok alone is far below it — so check each channel on its own, using revenue from your store.
Read the articleROAS only counts the sales an ad brought back, not what each order leaves you after product cost, fees, shipping and returns. Your break-even ROAS is 100 divided by what you keep from every 100 in sales — keep 40 and you need 2.5x, keep 25 and you need 4x — so a campaign at 3x can still lose money.
Read the articleROAS falls for one of two reasons: each click got more expensive (CPM, CTR or CPC moved in the ad platform), or each visitor became worth less after the click (conversion rate or average order value fell in your store). Split the funnel at the click, find which side moved, and fix the bigger leak first — one change at a time.
Read the article