What a TikTok report has to reconcile
The TikTok measurement hub sets the rules this tutorial turns into a routine. TikTok measures paid and organic results in two different places. Ads Manager counts paid results from the events the pixel and the Events API send, credited to ads under an attribution setting the advertiser chose, and reports them through custom columns, breakdowns and saved reports. The account's analytics report organic results per post and for the account, including where views came from: the For You feed, Following, the profile, search, a sound or a hashtag. Neither agrees with a web analytics tool, and the hub explains why the gap is expected rather than a fault.
Four overlaps cause most of the errors in TikTok reports, and every step below exists to handle one of them.
| Overlap | Where it hides | How the report handles it |
|---|---|---|
| Spark Ad engagement | Counted as paid engagement in Ads Manager and included in the post's organic totals | Subtracted from organic totals before the organic trend is read |
| View-through conversions | Counted by TikTok after a view without a click; often credited elsewhere by analytics | Kept on a separate line and never added to a total |
| Attribution setting | A changed setting changes the count with no change in results | Recorded with its date; months on different settings are not compared directly |
| TikTok against analytics and the order system | Each counts a different unit under different rules | One system named as the owner of each figure |
The organic TikTok hub lists the questions organic analytics can answer, and the Spark Ads guide explains the paid engagement columns. The social media reporting guide covers organic reporting across platforms, and the performance reporting guide covers paid reporting across channels. This page is the TikTok month in particular, where the two meet.
TikTok's Ads Help Center and the Creator Portal, listed in the sources, are the reference for the screens, the column names and the analytics views. They change from time to time, so the steps describe what to pull and why rather than where to click.
Before you start
- Access to the account's analytics through a Business Account, as the organic hub describes, for whoever pulls the organic figures.
- Reporting access to the ad account in Ads Manager, ideally with a saved custom report that already holds the columns you use every month.
- A web analytics property that receives TikTok clicks tagged with campaign parameters (UTMURL tags identifying traffic source.) mirroring campaign, ad group and ad names; the UTM and naming conventions guide sets out the rules.
- The business's own results for the month: orders and revenue from the shop, or leads and their outcomes from the CRM.
- A working pixel and Events API setup, deduplicated with a shared event identifier; the pixel and Events API tutorial installs and verifies it.
- Last month's workbook, or, for a first report, as many earlier months as the systems still hold.
- The targets the business set for TikTok this quarter, so the report has something of its own to measure against.
The steps
Write the questions the report answers
Before pulling a figure, write down the few questions the reader needs answered this month. Most TikTok reports need three or four: is organic content reaching new people, is paid delivery buying results at a cost the business accepts, what did TikTok add to orders or leads, and which creative should get more budget or a follow-up. Tie each question to one key performance indicator (KPIA metric tied directly to an objective.) and a source. Questions that no decision depends on produce vanity metrics: follower count, total views and total likes are worth a line only when they explain one of the indicators. The questions also set the order of the report, which should follow the reader's decisions, not the order the systems export data in.
Fix the period, the attribution setting and the owners of each figure
Use calendar months, in the same time zone across every system, and note any day that is not comparable, such as a sale, a launch or a tracking break. Record the attribution setting the ad groups used, with the date it was last changed, because the measurement hub is plain that the setting changes the numbers. AttributionAssigning conversion credit to touchpoints. is a choice, so write it at the top of the report. Then name the owner of each figure, as the hub recommends: TikTok for in-platform comparisons and optimisation, analytics for comparing TikTok with other channels, and the order or lead system for revenue and real outcomes. Write the owners into the workbook's first tab, so nobody has to argue about them again next month.
Pull organic figures by post and by traffic source
From the account's analytics, take the month's overview figures and then the per-post figures for every video published or still earning views: views, average watch time, the share watched to the end, profile views, follows, shares and the split of views by traffic source. The traffic source split matters most, because it says whether the feed showed a post to people beyond its followers and whether anyone found it through search. Export per post rather than only the account totals; an average of very different videos hides the post that did the work. Mark each post with its format or series, so the comparison step can read formats rather than individual videos.
Pull paid figures by campaign and creative
From Ads Manager, run the saved custom report by campaign, then by ad, with spend, impressions, reach, frequency, clicks, the video watch columns, conversions and cost per result. Add the breakdowns the questions need, by time, placement, age or country, rather than every breakdown on offer. Keep click-through and view-through conversions in separate columns where the report allows it. Label each ad with its creative concept and its identity, a Spark post or an uploaded video, as the campaign naming convention should already do. Where the business sells online, put revenue beside spend and work out the return on ad spend for each campaign with the ROAS calculator, using TikTok's attributed revenue for in-platform comparison only.
Separate Spark Ad engagement from organic totals
For every post that ran as a Spark Ad during the month, take the paid engagement columns from Ads Manager, the paid likes, comments, shares, follows and profile visits the ad earned, and set them beside the post's totals in organic analytics. The Spark Ads guide explains that those totals include the paid engagement, so the same like appears in both tools. Subtract the paid figures to get an organic-only line for the post, and report the paid engagement once, in the paid section. Where the analytics do not separate paid views from organic views for a post, note it and read that post's organic figures for the days before and after the flight instead. Without this step a month with heavy Spark spend looks like an organic breakthrough.
Add analytics sessions and the business's own results
From the web analytics property, take sessions, engaged sessions and conversions from TikTok traffic, by campaign, using the campaign parameters. Compare TikTok's click-through conversions with these sessions and conversions, like with like, rather than TikTok's total with the analytics total. Then add the order or lead system's figures for the month: total orders and revenue, or leads and the outcomes the CRM recorded. Put TikTok's view-through conversions on their own line beside them. Where the figures diverge more than usual, the platform versus analytics discrepancies guide gives the method for finding out why, and the report should say which cause was found rather than averaging the numbers.
Compare against the account's own months
Compare each indicator with the previous month and with the same month last year where the account is old enough, aligned for weekdays, sales and launches as the guide to comparing periods explains. Flag any month on a different attribution setting instead of comparing it directly. Read the count behind every rate: a cost per result built on a few dozen conversions, or a completion share from a handful of posts, moves for reasons that have nothing to do with the work, and the guide to reading small numbers shows how to put a range around it. Do not compare with rates quoted from other accounts; the report's standard is the business's own history and the targets it set.
Write what happened, why it matters and what changes
Write the commentary with the what, so what, now what framework: what happened against the questions and the targets, why it matters and what probably caused it, and what will change next month. Keep the middle step honest: a cause the data cannot support is written as a hypothesis with the check that would confirm it. End with a short list of decisions, each with an owner and a date, such as a creative series to expand, a campaign to pause or a test to run. A report that ends without decisions has described the month without using it. When the question is whether TikTok caused results beyond what last-click counting shows, the answer needs a test rather than a report, and the guide to measuring what TikTok adds beyond last click explains which evidence answers it.
Keep the workbook for next month
Save the month's exports, the joined tables and the commentary in one workbook the business owns, with a tab for each source and a tab for the joined view. Record anything that changes how next month will read: a new pixel event, a changed attribution setting, a new campaign type, a paused series. Next month starts by checking last month's decisions, so the decision list belongs at the top of the next report. Teams that report several channels fold the TikTok page into the monthly marketing report template and produce it through the monthly reporting pack SOP; the monthly paid media review checklist covers tracking health and pacing across all paid channels before the commentary is written.
The report on a page
A monthly TikTok report fits on one page when each section answers one of the questions from the first step and names its source. The layout below is a starting point; drop any section no decision depends on.
| Section | What it shows | Owner of the figures |
|---|---|---|
| Decisions | Last month's decisions with their outcome; this month's decisions with owners and dates | The report author |
| Business results | Orders and revenue, or leads and their outcomes, for the month against target | Order or lead system |
| Paid delivery | Spend, reach, frequency, click-through conversions and cost per result by campaign | Ads Manager |
| View-through line | View-through conversions, shown separately and never added to a total | Ads Manager |
| Organic reach | Views by traffic source and watch time by format, with Spark Ad engagement removed | Account analytics |
| Spark Ads | Paid engagement and results for each promoted post | Ads Manager |
| TikTok traffic on site | Sessions, engaged sessions and conversions from tagged TikTok clicks | Web analytics |
| Creative | The ads and posts that carried the month, by concept and identity | Ads Manager and account analytics |
| Changes and caveats | Attribution setting with its date, tracking changes, days not comparable | The report author |
Testing which creative works is a separate job from reporting it; the guide to testing TikTok ad creative says which method answers which question, the ad creative hook test designs one controlled comparison, and the report records what such tests found. A report that keeps showing the same gaps between TikTok and the shop's records is a reason to run the TikTok Ads account audit scorecard, which scores events, structure and reconciliation together and ranks the fixes.
A pet food brand's month (illustrative)
Vrandelup Pet Food reads a month where paid and organic overlap
Vrandelup Pet Food, an invented brand selling dog and cat food through its own website, reports TikTok each month on one page. The account's analytics show 410,000 organic views for the month against 520,000 the month before, but the top post, a feeding-time video, shows 96,000 views and 3,400 likes in its totals. It ran as a Spark Ad for the last 12 days of the month, and Ads Manager reports 2,900 paid likes and 610 paid follows on it.
| Line | Figure | Owner |
|---|---|---|
| Organic views, all posts | 410,000 (previous month 520,000) | Account analytics |
| Top post likes, total / paid / organic only | 3,400 / 2,900 / 500 | Analytics less Ads Manager |
| Paid spend | £4,200 | Ads Manager |
| Click-through purchases | 118 | Ads Manager, 7-day click setting unchanged since March |
| View-through purchases | 64, separate line | Ads Manager |
| Sessions from tagged TikTok clicks | 5,300, with 81 orders | Web analytics |
| Shop orders, all channels | 1,940 (previous month 1,870) | Order system |
Read with the paid engagement removed, the top post earned only 500 organic likes, and organic views fell for the second month across every format except short how-to clips. The 64 view-through purchases sit beside the shop's 1,940 orders and are not added to the 118 click-through purchases. The report ends with three decisions: the social lead commissions four new how-to clips before the next report; the paid media manager moves £800 from the weakest ad to two new Spark posts; and the founder approves a holdout test before the next quarter to see what TikTok adds beyond click-through orders.
Common mistakes
- Adding Spark Ad engagement twice. Paid likes and follows are in Ads Manager and in the post's organic totals; remove them from one before reading the organic trend.
- Folding view-through conversions into the total. Keep them on their own line, next to the business's own orders, and never add them to click-through conversions.
- Comparing months on different attribution settings. Record the setting and its date every month, and flag the months that cannot be compared directly.
- Comparing with someone else's rates. The account's own months and its own targets are the only fair standard.
- Reporting account totals only. One strong post can hide a format that stopped working; read by post and by format.
- Reading a rate without its count. A cost per result from a handful of conversions is noise until it has a range around it.
- Ending on charts. Every report ends with decisions, owners and dates, and the next report starts by checking them.
