The TikTok Ads Scorecard
This scorecard tracks the numbers that decide whether TikTok Ads pay off: whether people stop scrolling, whether they keep watching, how fast new creative ships, and whether the results the platform reports actually show up in your business. The media buyer runs the daily rows, the marketing lead runs the weekly rows, and the owner reads one number a month. Creative comes from your partners. We measure it.
Why this scorecard exists
Why a generic scorecard breaks on TikTok
Most scorecards were built for search and then copied over to social. That works badly on TikTok because the thing that drives cost on TikTok is not a keyword or an audience. It is the first three seconds of a video. A scorecard that only shows spend, clicks, and cost per result tells you something went wrong about two weeks after the creative stopped working. By then the budget is gone.
TikTok also burns through creative faster than any other platform we manage. A video that wins in week one is often losing by week three, and the ad manager will keep spending on it because it was the winner last time it checked. That makes creative supply a measurable input, not a vibe. When creative comes from a partner, an in-house team, or creators through Spark Ads, the scorecard has to show how many new pieces went live, which ones held attention, and how long since the last winner. Otherwise the conversation becomes about taste instead of throughput.
When the scorecard is live, the weekly meeting changes shape. Hook rate and hold rate get read before anyone talks about cost, because they explain the cost. The creative partner gets a brief that says which concept to make more of, based on numbers rather than opinion. The pixel and Events API get checked every day, so a broken event never spends a full month teaching the algorithm the wrong thing. The owner sees cost per result against their own history and knows whether the channel is earning its budget.
The scorecard
11 numbers. Every one has an owner.
| Metric | What it means | How to set the target | Owner | Cadence | Red flag |
|---|---|---|---|---|---|
| Hook rate TikTok charges you for the impression whether or not anyone watches, so this is the first number that explains every cost number below it. | The share of impressions that turn into 3-second video views, measured per ad and rolled up by creative concept. | Pull the trailing 8 weeks of hook rate for every ad that spent meaningfully, take the median, and set the floor there. New creative has to beat the median of the current winners to keep spend, and the floor moves up each quarter as the library improves. | Media buyer | Daily | Any ad that falls below the account floor on two consecutive days with real spend gets its budget capped and goes into the weekly creative review. |
| Hold rate A strong hook with a weak hold means the opening promised something the video did not deliver, and that shows up as expensive clicks that do not convert. | The share of 3-second viewers who are still watching at the ad's midpoint, or at the point where the offer is stated. | Measure hold at the same timestamp for every video in a concept so the comparison is fair, then use the median of your current winners as the bar. If your videos vary a lot in length, set the checkpoint as a percentage of runtime instead of a fixed second. | Media buyer | Weekly | A concept whose hold rate drops below the account median for two straight weeks while hook rate holds means the body of the video is the problem, and the brief to the creative partner says so. |
| New creatives launched TikTok fatigues creative faster than any other platform we run, so the number of fresh pieces entering the account is the leading indicator for next month's cost. | The count of net-new videos, not edits or re-cuts, that went live with spend in the week. | Look at how many weeks your past winners stayed above the hook floor before fading, then set the launch count so a new winner is always in testing before the current one dies. If winners last about three weeks, you need new candidates every week. | Marketing lead | Weekly | Zero new creatives for two consecutive weeks triggers a call with the creative partner that day, not at the next monthly check-in. |
| Days since last new winner Launch count measures volume, but this row measures whether the volume is any good, and it is the number the creative partner should be judged on. | The number of days since a newly launched creative beat the current top ad on cost per result at comparable spend. | Take the average gap between winners over the last two quarters and set the alert at that gap. As the partner learns what works, the target tightens. | Marketing lead | Weekly | When the gap passes the target, the next brief to the creative partner is a formal one with the hook and hold data attached, and the concept that has never won gets retired. |
| Top ad spend concentration TikTok's delivery piles budget onto whichever ad won first, which is efficient right up until that ad fatigues and takes the whole account down with it. | The share of weekly spend going to the single best-performing ad. | Check where concentration sat in the weeks before past winners faded, and set the ceiling just below that. The point is to keep enough spend on challengers to find the next winner while the current one still works. | Media buyer | Weekly | Concentration above the ceiling with a rising cost per result on that ad means the buyer forces spend onto the next two challengers before the week is out. |
| Cost per result vs trailing median Comparing against your own history strips out seasonality, platform-wide auction swings, and whatever a vendor's benchmark deck claims. | This week's cost per primary result compared to the trailing 8-week median for the same result, in the same campaign objective. | Start with the trailing 8-week median as the line. Once the creative pipeline is steady, tighten the target by a set percentage each quarter and hold the buyer to it. | Media buyer | Daily | Three consecutive days above the median with spend on pace starts a creative and placement review before the weekly meeting, and above it by more than the agreed band pauses the worst ad group. |
| Spark Ads share of spend Spark Ads carry the comments, likes, and creator handle with them, and in most accounts we measure they behave differently enough from dark posts that they need their own line. | The percentage of weekly spend running through Spark Ads, which boost organic posts from your account or a creator's account, rather than standard in-feed ads. | Split the last quarter's results between Spark and non-Spark, compare cost per result and hold rate for each, and set the share to favor whichever delivered. Revisit the split each month rather than letting it drift with whatever creative happened to arrive. | Marketing lead | Weekly | Spark share moving more than the agreed band from target in a single week without a decision behind it means someone is launching whatever is available, and the buyer explains why. |
| Landing page conversion rate TikTok traffic is nearly all mobile and arrives mid-scroll, so a page that converts fine from search can quietly waste every dollar this channel sends it. | Conversions divided by sessions for traffic that arrived from TikTok, measured in your analytics tool rather than in TikTok Ads Manager. | Use your own TikTok-sourced conversion rate over the trailing quarter as the baseline and treat any test as a lift against that. Do not compare it to your search landing pages, which get a different visitor with a different intent. | Marketing lead | Weekly | Two consecutive weeks below baseline while click volume holds means the page changed or broke, and the marketing lead loads it on a phone before the meeting ends. |
| Event health TikTok's bidding learns from these events, so a broken or double-counted event does not just hide the truth, it teaches the algorithm to buy the wrong thing. | Whether the pixel and Events API fired the expected number of events, deduplicated correctly, and matched at the rate they did last week. | Set the expected event count from the ratio of platform-reported events to site sessions over the trailing 30 days. Any day outside the normal band of that ratio is a tracking problem until proven otherwise. | Marketing ops lead | Daily | Event count outside the band, or a dedup rate that moves, gets checked in the Events Manager the same day and no optimization change is made until tracking is confirmed. |
| Spend pacing TikTok can overspend or underdeliver by large amounts when creative fatigues, and pacing is the earliest number that shows the account's delivery is off. | Month-to-date spend as a share of the month's budget compared to the share of the month elapsed. | Pace to the calendar unless your own history shows a clear reason not to, such as a weekday pattern in results. Set the tolerance band from how far pacing has drifted in normal months. | Media buyer | Daily | Pacing outside the band for three days gets a budget adjustment that day, and underdelivery paired with a falling hook rate means the problem is creative, not budget. |
| Results in the CRM Platform-reported results include modeled and view-through conversions, and the owner should know how many of them the business can find. | Leads, orders, or appointments that TikTok reported as results and that actually appear in your CRM or order system, matched by click ID or contact. | Set the match rate from your own trailing quarter and watch the trend, not the absolute number. A falling match rate usually means the tracking or the form changed. | Owner | Monthly | Match rate falling below the trailing quarter for a full month starts an attribution review before any budget increase is approved. |
Targets are set from your own history, never from someone else's benchmarks. The builder does the arithmetic once you plug in your numbers.
The weekly review
Twenty minutes, same order, every week.
// 01
Cost per result against your own median
Open with the one number the owner cares about: this week's cost per result against the trailing 8-week median. Say whether it is inside the band, above it, or below it. Nothing else gets discussed until everyone agrees on that number.
// 02
Hook, hold, and the creative pipeline
Look at hook rate and hold rate by concept, then at new creatives launched and days since last new winner. This is where the cost number gets explained. The media buyer speaks here, and the answer is always about which creative to scale and which to cut.
// 03
Tracking, pacing, and landing page
The marketing ops lead confirms event health was green every day this week. The buyer confirms pacing. The marketing lead reads the TikTok landing page conversion rate. If any of these are red, the creative discussion above may have been based on bad data.
// 04
The brief and the budget
The meeting ends with two decisions written down: what the creative partner makes next week, with the hook and hold numbers behind the request, and whether budget moves up, down, or stays. Each decision has a name next to it and a date it gets checked.
Where scorecards die
Mistakes we see constantly.
Reading cost before reading the hook
Cost per result is the last number in the chain, not the first. Teams that open the meeting with cost end up arguing about bids when the real problem is that nobody is watching past second three.
Counting edits as new creative
A new caption or a re-cut of last month's winner is not a new creative, and TikTok's delivery treats it about the same way. The launch count only counts videos with a new hook or a new concept.
Letting the winner eat the budget
TikTok will happily put most of the week's spend on one ad. That feels efficient until the ad fatigues and there is nothing tested and ready behind it.
Trusting Ads Manager conversions alone
Platform results include modeled and view-through conversions that your CRM may never see. Without a monthly match against your own records, the scorecard is reporting TikTok's opinion of TikTok.
Checking the pixel once a month
A broken event spends the whole month teaching the algorithm to optimize for nothing. Event health is a daily row because the cost of a bad day compounds.
Other scorecards
Same discipline, different numbers.
By platform
By business type
Want it set up live in your dashboards?
We install scorecards on real data, wire the owners and alerts, and coach the weekly review. Scorecards and accountability is the service.
Questions we get
Straight answers.
Do you make the TikTok creative?
No. Creative comes from your team, your creative partner, or creators through Spark Ads. We build the measurement around it, write the briefs from the data, and tell you which concepts earned more spend and which should be retired.
What is a good hook rate for TikTok Ads?
We do not publish a number, because it changes by vertical, offer, and creative style. The useful bar is the median of your own current winners, and the useful question is whether new creative beats it. Set the floor from your trailing 8 weeks and move it as the library improves.
Why is Spark Ads share its own row?
Spark Ads run from a real account and carry its comments and social proof, so they tend to hold attention and convert differently from standard in-feed ads. Tracking the share of spend lets you decide the mix on purpose rather than by whatever creative showed up that week.
Who owns the daily rows if we do not have a full-time media buyer?
Someone still has to look every day, even if it takes five minutes. Hook rate, cost per result, pacing, and event health are the four daily rows, and if you do not have a buyer we can run them for you or coach the person who will.
How does this connect to the rest of our scorecard?
Results in the CRM is the row that links TikTok to the business scorecard, and it is the one the owner reads. Everything above it exists to explain that row. If you want to see how your current setup measures up, the Marketing Ops Scorecard self-assessment at /tools/marketing-ops-scorecard takes about ten minutes.
A scorecard is only as good as the system feeding it.
Start with a diagnostic call. We will tell you which of these numbers you can trust today and what it takes to trust the rest.