How to fix ads that get clicks but no sales

Your ads get clicks, Ads Manager says hundreds of people came, and your store has one sale or none. Most founders then change the ad, the audience and the budget all at once, or decide ads do not work for a D2C brand like theirs. This guide shows how to fix ads that get clicks but no sales: find the one step where shoppers drop out, fix that step, and test again on the same budget.

Do these three things

About 2 hours today, then about 1 hour a day and 30 minutes every Sunday. Judge a step only after enough people reach it: 100 clicks, 200 page views, 10 carts.

Show the three actions in fullHide the full list

1. 1Today: check your ad orders reach Shopify and Meta once each. Then put five numbers from your last ad test in one sheet: link clicks, landing page views, adds to cart, checkouts started and orders. Mark the first step that misses its line.
2. 2Every day: watch 20 recordings of shoppers in Microsoft Clarity (free), fix one thing at the step that leaks, and email every shopper who left a checkout to ask what stopped them, as part of winning back abandoned carts.
3. 3Every Sunday: count the five numbers again. Run your next ad test at the same limit only when every step meets its line. When it has spent its limit, raise 20% or stop.

Get help doing these

ChatGPT or Claude reads this guide and works through it with you.

Do this with ChatGPT(opens in a new tab)Do this with Claude(opens in a new tab)

1. Check your ad sales are being counted

Before you change any ad or page, check that the orders your ads bring show up once in Shopify and once in Meta. If tracking is broken, every number in this guide is wrong, and you would be fixing a store that may be fine.

Pick the dates of your last ad test. Then compare two lists:

1. In Shopify, open Analytics, then Reports, and open "Sales attributed to marketing". Count the orders tagged with your campaign. If your ads had no UTM parameters, open each order from those dates and click View conversion details under Conversion summary to see where the visit came from.
2. In Meta Ads Manager, add the Purchases column for the same dates.

Meta's attribution window means its count rarely matches Shopify's exactly. If every order Shopify shows from Facebook or Instagram also shows in Meta, and Meta shows no purchases you cannot find in Shopify, tracking works. Go to step 2.

If they do not match, fix tracking first. Place one real test order and check that Meta Events Manager shows exactly one Purchase, as in step 3 of how to run Meta ads for a Shopify store. Two purchases for one order usually means a second Meta pixel. Zero means the store is not connected properly. Setting up conversion tracking covers the full fix.

If you sell to the UK or the EEA, Shopify's cookie banner is on there. Shopify collects visitor data only after a shopper accepts, so Shopify will show fewer visits from your ads than Meta. That is expected. Count orders, not visits, in this step.

• Tool: Shopify Analytics and Meta Ads Manager. Meta Events Manager for the test order.
• Time: 20 minutes, or 50 if you need a test order.
• You will have: proof that each ad order shows once in Shopify and once in Meta.
• It worked if: the two lists match, order for order, for the dates of your last test.
• Common mistake: rebuilding the product page because Meta shows 0 purchases, when Shopify shows two real orders from the ads. Check the count before you check the store.

2. Count five numbers from click to order and find the leak

Put five numbers from your last ad test in one row: link clicks, landing page views, adds to cart, checkouts started and orders. Then compare each step with the one before it: each is a small conversion rate. The first step that misses its line is your leak, and you fix that one first.

Where each number comes from. Use the same ads and the same dates for all five.

• Link clicks and landing page views, from Meta Ads Manager. Click Columns, then Customize columns, and add both. A link click is a tap on your ad's link. A landing page view counts only when your page then loads, so the gap between them is people who left before they saw your store.
• Adds to cart, checkouts started and orders, from Shopify. Open Analytics, then Reports, and open "Conversion rate breakdown". In Filters, click the plus, choose UTM source, add the value facebook and click Apply. The report shows sessions with cart additions, sessions that reached checkout and sessions that completed checkout. These are visits, not products, so one shopper who adds two items counts once.

Shopify changed how it counts visits on 21 to 23 September 2026: it now leaves out bots. If your test ran before that, use its numbers, but compare later weeks only with weeks after that date.

The lines. These are this guide's lines for a store with a few hundred clicks. Each one is a bar for "this step is not broken", not a goal:

Why these: ad specialists put a normal click-to-page rate at 70 to 90%. Dynamic Yield's add-to-cart average over the last 12 months is about 6 items per 100 product page views, and about 10 for beauty and personal care. Baymard Institute's average across 50 studies is that about 7 in 10 carts are left, so about 3 in 10 become orders.

Before a step has had its "judge it after" number of people, do not judge it. If your last test stopped at 60 clicks, you do not know yet whether anything is broken. The stop line in how much to spend on ads is 3 times what an order leaves you, the same rule as in the ecommerce budget guide. If an order leaves you $20 to $25, that buys about 100 to 150 clicks at today's cost per click. In LocaliQ's 2026 benchmarks, a Facebook traffic ad click cost $0.60 on average and $0.50 for beauty and personal care.

Test: [name] · Dates: [from] to [to] · Spent: $[amount]
Link clicks (Meta): [number]
Landing page views (Meta): [number] → per 10 clicks: [LPV ÷ clicks × 10] · line 7
Adds to cart (Shopify, UTM source facebook): [number] → per 100 page views: [carts ÷ LPV × 100] · line 5
Checkouts started (Shopify): [number]
Orders (Shopify): [number] → per 10 carts: [orders ÷ carts × 10] · line 2
First step under its line: [click to page / page to cart / cart to order / none yet]

Paste the five numbers into ChatGPT or Claude and ask it to fill in the sheet and mark the first step under its line. Check each sum yourself.

• Tool: Meta Ads Manager, Shopify Analytics and Google Sheets.
• Time: 40 minutes.
• You will have: one row of five numbers, three rates and the first step that misses its line.
• It worked if: every number comes from a report, for the same ads and dates, and the leaking step is written down.
• Common mistake: judging the whole funnel on 30 clicks and 0 orders. At 30 clicks, 0 orders is normal even for a good store. Get to 100 clicks before you judge the first step.

3. Fix link clicks that never become landing page views

If fewer than 7 in 10 link clicks become landing page views, you are paying for taps from people who never see your store. Find where those clicks come from, make your page load fast on a phone, and put the campaign fixes into your next test.

Check these four things, in this order:

1. Where the clicks come from. In Ads Manager, open your campaign, click Breakdown, then By delivery, then Placement. Each place your ad showed gets its own row. Look for one placement with many link clicks, few landing page views and no orders. Audience Network, Meta's ads inside other apps, is a common source of accidental taps and bot clicks, mostly when a campaign aims for clicks. A high click-through rate with no orders there is a warning, not a win.
2. What the campaign aims for. Open the ad set. The goal should be Sales, with Website and Purchase as the conversion. A boosted post or a Traffic campaign finds people who click, not people who buy. How to run Meta ads for a Shopify store shows the setup.
3. Where the ads show. Under audience, the location should be only the countries you ship to.
4. How fast your page loads on a phone. Open PageSpeed Insights (free), paste the exact link your ad uses, and read the Mobile result. Largest Contentful Paint is how long the main picture or text takes to appear. Google calls 2.5 seconds or less good and over 4 seconds poor. If yours is slower, start with the first picture on the page: make the file smaller, or swap a video at the top for a photo. Remove apps you no longer use, because many add code to every page. Improving website speed covers more fixes. Shopify's own speed reports, under Online Store, need more visits than a new store has, so use PageSpeed Insights until they show numbers.

Also tap your own ad on your phone. Check the link opens the right product, in stock, with no error page and no redirect to the homepage.

When to change the campaign. Meta says changes to an ad set's audience or placements can restart its learning. So during a running test, turn ads off at the stop line and add spares, and change nothing else, as the budget rules say. Write the campaign fixes down and put them into your next test. Page speed and link fixes are on your store, so do them today.

To leave out a placement in the next test, open the ad set, choose manual placements, then Placement controls and Manage excluded placements. Since October 2025, Meta then shows a box, "Allow limited spend to excluded placements", which is ticked and lets Meta spend up to 5% of your budget on each placement you left out. Untick it if you want none. Leave out a placement only if the breakdown shows it took most of your clicks and brought no orders. Otherwise keep Meta's own choice, Advantage+ placements.

If you also run Google ads, open Campaigns, then Insights and reports, then Search terms. Add any search that could never buy your product, such as "free", "DIY" or "jobs", as a negative keyword, so your ad stops showing for it. Performance Max campaigns have their own search terms report. Google Shopping vs Performance Max covers which to use.

• Tool: Meta Ads Manager, PageSpeed Insights (free) and your phone.
• Time: 45 minutes.
• You will have: a list of fixes for the next test's settings, and a product page that loads in 2.5 seconds or less on a phone.
• It worked if: in the next test, at least 7 in 10 link clicks become landing page views after 100 clicks.
• Common mistake: changing placements and audience on day 3 of a running test. It restarts Meta's learning and mixes two tests into one, so you cannot tell what worked.

4. Make your product page say what your ad said

If people land but fewer than 5 in 100 add to cart, your page does not continue what the ad promised. Open each ad and the page it links to, side by side on your phone, and make the first screen of the page show the same product, the same promise and the price.

Shoppers who click an ad expect to see what the ad showed them. If they land on the homepage, a different product, or a page that opens with your brand story, most leave in a few seconds.

Ad name: [name]
What the ad promises, in its words: [one line]
Product and price in the ad: [product] · $[price or "not shown"]
Offer in the ad, if any: [discount, bundle, free shipping]
Link: [the exact URL the ad uses]On the page, on a phone, without scrolling:
Same product picture? [yes / no]
Same promise, in the same words? [yes / no]
Price, and the offer if any? [yes / no]
At least one review or star rating? [yes / no]
Delivery time and returns, near Add to cart? [yes / no]

Fix every "no", starting with the ad that had the most clicks:

• Send each ad to the product it shows, never the homepage or a collection page. The product page is your landing page for paid shoppers.
• Put the ad's promise at the top of the page, in the same words, above the fold. Your value proposition is the one sentence both should share. If the page cannot say it, change the ad instead: make new ad versions that promise what the page shows.
• Show the price and the full offer, the same as the ad. If the ad says "free shipping over $40", the page says it too.
• Show proof on the first screen: a star rating and one line from a real review, or a short clip of a customer or a small creator using the product. That is social proof, and user-generated content works well here.
• Answer the questions strangers ask in the first lines of the product description: what it is made of, the size, how long delivery takes and how returns work. How to make a new store look trustworthy lists what else cold shoppers check.

ChatGPT or Claude can draft the top lines from your ad's words and your reviews in a few minutes. You check every claim is true.

Then watch what shoppers do. Install Microsoft Clarity's own app from the Shopify App Store. It is free, and it records visits as videos and shows where people tap and how far they scroll. Every day, open Recordings, filter by Source to facebook (your UTM tag), and watch 20 recordings, or every one if fewer came. For each, write down where the shopper stopped: on the first screen, at the price, at the photos, at shipping. Fix one thing a day at the step that leaks, the one most shoppers stopped at. One change a day keeps each fix easy to check; a proper A/B test needs far more visitors than you have yet.

And ask people. While ads are off, keep doing the free work from the budget guide: 100 personal messages a day. In 20 of them, ask the person to open your product page for 5 seconds and tell you what it sells and what it costs. It is a two-question version of a customer interview. If more than 5 of 20 cannot say both, fix the first screen again.

• Tool: Shopify's theme editor, Microsoft Clarity (free) and ChatGPT or Claude.
• Time: 2 hours for the first pass, then 30 minutes a day for recordings and one fix.
• You will have: a filled check for every live ad, a page whose first screen repeats each ad's promise, and a daily note of where shoppers stop.
• It worked if: after 200 landing page views, at least 5 in 100 add to cart, and 15 of 20 people can say what the page sells and what it costs.
• Common mistake: buying a new theme or redesigning the whole store. The leak is usually one missing thing on the first screen: the promise, the price or the proof. If the page step still misses its line after these fixes, fix the whole store before your next test.

5. Fix add to cart but no purchase at checkout

If people add to cart but fewer than 2 in 10 carts become orders, something at checkout surprises them. Buy your own product on your phone as a stranger, remove each surprise you find, and email every shopper who left a checkout to ask what stopped them.

Buy as a stranger. Open your ad link on your phone in a private window, add the product and go through checkout to the payment page. Write down every surprise. In Baymard Institute's survey of US shoppers who left a checkout, the top reasons were extra costs such as shipping and tax (40%), slow delivery (20%), not trusting the site with a card (19%), having to make an account (18%) and a long checkout (17%).

Fix them in this order:

1. Show the shipping cost before checkout. Put the shipping price, or "free shipping over $[amount]", on the product page next to the price. A free-shipping line works well with a bigger order, such as a bundle.
2. Let people buy without an account. If checkout asks you to sign in or make an account, turn that requirement off in your Shopify checkout settings.
3. Turn on express pay buttons. Branded buttons such as Shop Pay, Apple Pay, Google Pay and PayPal show next to Add to cart once those ways to pay are on in Settings, then Payments. Shop Pay comes with Shopify Payments, at its normal rates.
4. Say when it arrives and how returns work on the product page and in the cart.

Ask every shopper who left. Shopify lists every checkout where a shopper gave an email and did not pay under Orders, then Abandoned checkouts. Every day, open each new one, click the shopper's email, and send a short personal note with one question. Send the recovery link only to that shopper, because anyone with it can finish the checkout. Follow up up to 3 times, until they answer. Their answers are the most useful feedback from early buyers you will get. AI drafts each note in 30 seconds; you add one true line about what they chose.

Subject: your [product] is still in your cartHi [first name],I'm [your name], I run [store] myself. I saw you got as far as checkout for the [product] and stopped. Was something wrong, or missing?One line back helps me more than you'd think. If it was the shipping cost, the delivery time or something on the page, tell me and I'll fix it.Your cart is here if you still want it: [recovery link][your name]

Also turn on Shopify's own reminder: open Apps, then Messaging, then Automations, and turn on "Abandoned checkout emails by Shopify". Choose who it goes to under "Send to". For the full email flow, see Klaviyo flows for a new store. Shoppers who leave earlier, at the cart, never give an email. To reach them, retargeting ads come later, once you have more visitors.

• Tool: your phone, Shopify (Orders, Settings and the Messaging app) and your email.
• Time: 1 hour to buy as a stranger and fix the surprises, then 15 minutes a day for the emails.
• You will have: a checkout with no surprise costs, guest checkout and express pay, and a log of every shopper's answer.
• It worked if: after 10 adds to cart, at least 2 become orders, and every shopper who left a checkout has had an email.
• Common mistake: adding a discount pop-up to stop people leaving. If the shipping cost is the surprise, a 10% code does not remove it, and it trains shoppers to wait for one. Show the cost up front instead. If you want a pop-up, use one that is not a discount. Abandoned carts covers the rest.

6. Every Sunday, count the five numbers again and plan the next test

Every Sunday, count the same five numbers for the week, write them under last week's, and decide on your next ad test. Run it at the same limit as the last one, and only when every step meets its line.

While your ads are off, count the page and checkout steps on all your visitors that week, from the 100 messages a day and your free posts on Instagram. Take the Shopify report without the UTM filter. Visitors who come from a message are warmer than ad visitors, so treat a line met on them as a first pass, not proof.

Then do the click price sum. It tells you whether ads can pay for your store at all, even with nothing broken. It starts from what one order leaves you, your break-even customer acquisition cost:

What one order leaves you (break-even cost): $[X]
Orders per 100 link clicks: [orders ÷ clicks × 100]
Most one click can cost: $[X] × [orders per 100 clicks] ÷ 100 = $[most per click]
What a click costs you now (Meta, cost per link click): $[CPC]
Orders per 100 clicks you need at that price: [CPC ÷ X × 100]

Meeting the three lines means no step is broken. It does not mean ads pay. The last line of the sum is the number your next test must reach. If the most one click can cost is below what you pay, you need more orders per click or more money left from each order: a higher price, a bundle that raises the average order, or repeat orders. Your break-even ROAS is the same check from the other side, measured in ROAS.

Then decide:

• Every step meets its line: run your next test at the same limit as the last one, not bigger (or at 10 times your break-even cost if your last test had no limit), with the campaign fixes from step 3 and your best spares. Testing ad creatives side by side comes once tests pay back. Set it up as in how to run Meta ads for a Shopify store. When it has spent its limit, follow the rule there: at or under your break-even cost per order, run the next one 20% bigger; over it, stop and come back to step 2 with the new five numbers. Compare its orders per 100 clicks with the number the sum said you need.
• A step still misses its line: keep ads at $0 and keep fixing that step, one thing a day.
• Two tests in a row miss with every step at its line: the offer is the problem, not the ads. Keep ads at $0, go back to why no one is buying, and grow with the free channels in how to get customers online for free. Choosing your channels shows where ads fit later.

When a new test runs, check every ad each day against its stop line, as before. Once a month, step back for a monthly marketing review.

• Tool: Google Sheets, Shopify Analytics and Meta Ads Manager.
• Time: 30 minutes every Sunday.
• You will have: one row of five numbers a week, the click price sum, and a decision for the next test.
• It worked if: the decision follows the rule above, and the next test uses the same limit as the last.
• Common mistake: raising the budget "to get more data" before the leak is fixed. More money through a broken step buys more clicks that do not buy.

Where Scout7 helps

Scout7 helps with the ad parts of steps 3, 4 and 6. Campaigns run in the Scout7 app, not from a chat with Claude or ChatGPT yet.

Step 3. Scout7's campaign monitoring reports spend, impressions, clicks and click-through rate for each campaign, and every 12 hours proposes actions, such as which creative to stop. You approve or reject each one. Landing page views and placements still come from Ads Manager, and orders from Shopify.

Step 4. When the fix is a new ad that promises what your page shows, Scout7's Image Studio makes image posts with captions in your brand's voice, and Video Studio writes a script, films it scene by scene and adds captions, in portrait 9:16. You direct the scenes and approve each one.

Step 6. For the next test, Scout7 connects your Meta ad account, drafts the campaign with an audience, an estimated reach and the budget you set, and walks it through creative review, targeting and budget, including a lifetime budget. Enter the same limit as your last test. You approve every creative and the spend; nothing launches without your approval.

Your checklist to fix ads that get clicks but no sales

0 of 13 done

Work through this list with an AI assistant

It drafts the work and keeps track of what is done.

Do this with ChatGPT(opens in a new tab)Do this with Claude(opens in a new tab)

Step 1 · Check your ad sales are being counted

• Compare the orders Shopify shows from your ads with the purchases Meta shows for the same dates.
• If they do not match, place one real test order and fix tracking until Meta shows exactly one purchase.

Step 2 · Count five numbers and find the leak

• Put the five numbers from your last test in one sheet and mark the first step under its line.

Step 3 · Fix link clicks that never land

• Break down your last campaign by placement and write the fixes for the next test.
• Make the product page your ad links to load in 2.5 seconds or less on a phone.

Step 4 · Make your product page say what your ad said

• Fill in the ad-to-page check for every live ad and fix every “no”.
• Every dayWatch 20 recordings in Microsoft Clarity and fix one thing at the step that leaks.
• Every dayWhile ads are off, send 100 personal messages and ask 20 people to read your product page for 5 seconds.

Step 5 · Fix add to cart but no purchase

• Buy your product on your phone as a stranger and remove every surprise you find.
• Every dayEmail every shopper who left a checkout and ask what stopped them.

Step 6 · Every Sunday, count again and plan the next test

• Every SundayCount the five numbers for the week and compare each step with its line.
• Every SundayDo the click price sum.
• Every SundayDecide the next test by the rule: run it at the same limit only when every step meets its line.

Copy listDownloadReset