How to Analyse a TikTok Ads Campaign: The Metrics That Matter
A TikTok report gives you dozens of numbers, and most of them are true. The difficulty is choosing the ones that explain a decision. When a campaign disappoints, you need to know where the buyers were lost: before they watched, after they watched, on the store, or inside the tracking itself.
Why single metrics mislead
Three habits produce wrong conclusions:
- Judging by one number. A low cost per click looks good until you see that nobody buys. A high view count looks good until you see the viewers never click.
- Comparing against a borrowed benchmark. Averages from other industries and markets say little about your margin. The only benchmark that matters is your own break-even; see How to Calculate Your Advertising Break-Even Point Before Scaling.
- Trusting platform labels without checking definitions. View metrics in particular are defined differently across tools. TikTok's reporting documentation notes, for instance, that in Ads Manager the 2-second and 6-second view metrics are absolute counts rather than percentages, while some reports show them as rates. Check how your account defines each metric before comparing campaigns.
The metric ladder
Read the campaign as a sequence of steps, and find the weakest step. Each step answers one question.
- Impressions to views. Does the opening hold attention? Look at early-view counts relative to impressions.
- Views to clicks. Does the ad make people want the next step? Look at click-through rate.
- Clicks to add to cart. Does the store match the promise of the ad?
- Add to cart to purchase. Does checkout, price, delivery or trust stop people?
- Purchase to profit. Does the order leave margin after product, shipping and failed parcels? See Contribution Margin per Order.
Impressions to 2-second views
38,000 of 100,000
38%2-second to 6-second views
12,000 of 38,000
31.6%Clicks
900 of 100,000 impressions
0.9%Add to cart
90 of 900 clicks
10%Purchases
27 of 90 add to cart
30%
100,000 impressions, spend 1,350 SAR, average order 180 SAR.
A worked example
Illustrative example, not real results. A campaign spends 1,350 SAR and reaches 100,000 impressions. It records 900 clicks, 90 add-to-carts and 27 purchases, at an average order value of 180 SAR.
- Click-through rate: 900 / 100,000 = 0.9%.
- Click to purchase rate: 27 / 900 = 3%.
- Cost per click: 1,350 / 900 = 1.50 SAR.
- Cost per purchase: 1,350 / 27 = 50 SAR.
- Revenue: 27 x 180 = 4,860 SAR; revenue per riyal spent: 4,860 / 1,350 = 3.6.
- Cost per thousand impressions: 1,350 / 100,000 x 1,000 = 13.50 SAR.
Now read it. Revenue per riyal of 3.6 is not profit. If the margin per order before ads were 45% of 180, that is 81 SAR, and the 50 SAR cost per purchase leaves 31 SAR before failed deliveries. If the margin were 25%, it is 45 SAR and the same campaign loses 5 SAR per order. The identical campaign is healthy or loss-making depending on your cost structure, which is why the ladder ends at profit and not at the platform's return figure.
Which step is weakest? In the example, the largest drop is between add-to-cart and purchase (30%). That points to checkout, delivery terms, price surprises or trust, not to the creative. Changing the ad would address the wrong layer.
Each bar shows the share kept from the previous step.
Diagnosis by where the ladder breaks
- Few early views. The hook or the first frame is weak. Test openings; see TikTok Ad Creative: The First Three Seconds and What to Test.
- Good views, few clicks. The ad entertains but does not sell. Check the offer and the call to action.
- Clicks but no add-to-cart. The landing page or product does not match the promise, or the page loads slowly on mobile.
- Add-to-cart but no purchase. Look at delivery cost shown late, payment options, trust signals, and for cash on delivery stores, confirmation friction.
- Purchases but no profit. The issue is margin or return rate, not advertising.
What the data cannot tell you
- Attribution is partial. A viewer may see the ad on TikTok and buy later through a search or a direct visit. Platform-reported purchases and your store's orders will not match exactly; compare trends, not single days.
- Tracking errors look like performance errors. If the pixel misses events or counts a purchase twice, the ladder lies. Verify with TikTok Pixel Setup for E-commerce and Duplicate Purchase Events: Causes and How Deduplication Works.
- Small numbers are noisy. 27 purchases is a sample, not proof. Do not rewrite the strategy on a handful of orders.
- A late-stage metric can lag. Delivered and kept orders arrive days after the click.
When the advice changes
If the product is a high-consideration purchase, the ladder is slower and you should add a step for leads or messages. If you use a messaging or form flow instead of a store checkout, replace steps 3 and 4 with your own stages. If the account is new and volumes are small, judge the first stages only and wait before judging purchases.
Your next action
Put your last campaign into the ladder with real numbers: impressions, early views, clicks, add-to-carts, purchases, spend. Mark the largest drop. Fix that step only, and measure again. If the numbers disagree between TikTok, your store and your analytics, a Full Ads & Tracking Audit finds out which is wrong before you spend more. After you have a stable pattern, Scaling TikTok Ads: Creative Refresh Cycles and Budget Steps covers the next phase.