# How to Check a TikTok Engagement-Rate Spreadsheet for Formula Errors

To check your TikTok engagement-rate spreadsheet, write the intended formula first. Trace each count to its source cell. Check missing values and units. Then calculate a few rows by hand and compare them with the sheet. Keep a correction log. A percentage can look reasonable even when the formula uses the wrong denominator or applies percentage formatting twice.

Start with the supplied spreadsheet and its source records. You do not need a new data collection run to find a wrong cell reference or a display error. Collect fresh public counts only when the review needs a new observation. A fresh count does not establish what an older cell should have contained.

This guide checks the sheet’s arithmetic and source references. It does not set a universal engagement formula. It also does not turn action counts into a percentage of unique people.

## 1. Write the intended rule before checking the result

First, state the question that the sheet must answer. Then write the formula in words.

The numerator is the amount above the division line. The denominator is the amount below it. A ratio is the result of dividing one amount by another. A percentage expresses that ratio per hundred.

For this guide’s worked example, the written rule is:

`(likes + comments + shares) / views`

The sheet must display the result as a percentage. This is the rule for the example. It is not a rule for every TikTok report.

Record these details beside the sheet:

1. The question the rate must answer.
2. The action counts included in the numerator.
3. The counter used as the denominator.
4. The time at which the counts were collected.
5. The rule for missing or rounded counts.
6. Whether the stored result is a ratio or a percentage-point value.

A percentage-point value stores the percentage number itself. For example, the number 10 can represent 10 percentage points. A stored ratio of 0.1 represents the same proportion, but the display must handle it differently.

The denominator needs special care. One supplied [public reader question about conflicting formulas, published May 12, 2025](https://www.reddit.com/r/influencermarketing/comments/1kl020c/engagement_rate_help_brand_deal_advice_food/) asks whether to divide by followers or views. That is a definition question before it is a spreadsheet question. The record shows one reader’s question; it does not establish how common the problem is.

Use the existing [guide to defining and comparing social view ratios](/en/articles/social-view-metric-comparability/) when you need to settle what the denominator means. Then return to this audit, which is a check against the declared rule. Do not choose a denominator because it produces a more attractive result.

## 2. Trace each source field to its spreadsheet cell

An observation is one record of a post at a stated collection time. Check that each formula uses the intended observation.

A cell reference identifies the spreadsheet cell that supplies a value. A formula can perform correct arithmetic with an incorrect reference. For example, it can use a follower count where the written rule requires views.

Make a field-to-cell map. A field is a named value in the source record. The map connects each field to its destination in the sheet.

| Save this detail | What to check |
| --- | --- |
| Post identity | The source and sheet refer to the same post |
| Post URL | The link leads to the intended post |
| Collection time | The counts belong to the intended observation |
| Source field | The count has the meaning required by the rule |
| Source value | The sheet preserves the supplied value |
| Destination cell | The formula uses the intended cell |
| Availability | The count is present or unavailable |
| Precision | The count is exact, rounded, or uncertain |
| Evidence reference | Another reviewer can locate the source record |

Keep the source value beside any corrected value. Do not overwrite the only record of what the sheet received.

Check identity before comparing numbers. A count from another post can still look plausible. Also check time. A current follower count must not silently replace an older source value during an arithmetic review.

If you use owner-only records, identify them as such. These are records available to the account owner rather than through the public page. Keep their meaning and reporting period visible. Do not assume a public counter and an owner-only record measure the same thing merely because their labels look similar.

## 3. Use the Actor only when you need a public source observation

The manual spreadsheet review comes first. The [TikTok Metrics Scraper](https://apify.com/chronometrica/tiktok-metrics-scraper) can supply documented public source fields when you need a new observation. An Actor is a hosted data collection program. It does not perform this formula audit or produce your correction log.

The supplied public contract for build 0.1.36 documents direct video input through `startUrls`, with `mode` set to `knownUrls`. This small input uses that documented structure:

```json
{
  "maxItems": 1,
  "mode": "knownUrls",
  "startUrls": [
    "https://www.tiktok.com/@mrbeast/video/7654638524729216287"
  ]
}
```

3. Use the Actor only when you need a public source observation

Replace the entire sample URL with the intended public video URL. The sample address is not evidence for your client’s spreadsheet. `maxItems` limits the number of output rows; here, the limit is one. A one-post sample is not a complete account archive.

This is a proposed collection input. No run was made for this guide. The supplied [public build contract, checked October 25, 2026](https://api.apify.com/v2/acts/chronometrica~tiktok-metrics-scraper/builds/default) establishes the documented input and output fields. It does not establish a successful collection result.

Useful documented fields include:

- `postId` and `postUrl` for post identity.
- `scrapedAt` for collection time.
- `likes`, `comments`, `shares`, and `views` for this example’s formula.
- `postMetricStatus` for each metric’s availability and precision.
- `status` and `statusReason` for the record’s status and explanation.
- `profileFollowersCount` when a declared rule needs a follower count.

A metric is a measured count or value. Precision describes how closely a reported value represents the underlying count. Inside `postMetricStatus`, the documented `status` values are `present`, `unavailable`, and `not_attempted`. The documented `precision` values are `exact`, `display_rounded`, `likely_rounded`, `unavailable`, and `unknown`.

Inspect these fields rather than judging quality from the number alone. The public contract allows missing metric values as `null`, which means no value is supplied. It also documents best-effort public profile enrichment. Public counts do not establish private reach or unique engaged people.

A Task is a saved input preset. The supplied evidence confirms that the [TikTok Creator Analytics Task page was publicly present on October 25, 2026](https://apify.com/chronometrica/tiktok-metrics-scraper/examples/tiktok-creator-analytics). Its saved settings were not inspected through an authenticated account, and the Task was not run. Do not assume its saved inputs match your review.

## 4. Check every required numerator component

Check each action count before adding the counts together. The declared formula requires likes, comments, and shares. All three must have usable values for a complete numerator.

A missing value is not zero. Zero says that the recorded count is zero. A blank or `null` source value does not provide that count.

A spreadsheet sum can ignore an empty cell and still return a number. That behavior does not prove that the required numerator is complete. Check the component cells, not just the total.

Use a visible state beside the result:

| Source condition | Treatment in the sheet |
| --- | --- |
| All required counts are present and exact | Calculate after checking the denominator |
| A required count is missing | Mark the required result unavailable |
| A required count is rounded | Preserve that limit and label a calculated result approximate |
| A count’s precision is unknown | Keep the uncertainty visible |
| A recorded count is zero | Preserve zero as a supplied value |

These are proposed worksheet rules. They are not additional Actor output fields.

Do not add saves or reposts merely because those fields are available. That would change the declared formula. If the client needs another action measure, define it separately. The [guide to using shares and saves for content decisions](/en/articles/tiktok-share-save-decisions/) covers that separate choice.

Keep units clear. Likes, comments, and shares are action counts. Their sum is not a count of unique people. This arithmetic review cannot change that limit.

## 5. Check the denominator, value type, and stored units

Confirm the denominator’s source field before checking its amount. Under this guide’s example rule, the denominator must be `views`. A follower count fails that rule even when the division produces a reasonable percentage.

Then check these conditions:

1. Missing denominator: Mark the result unavailable.
2. Zero denominator: Mark the ratio undefined because division by zero has no defined result.
3. Text-formatted count: Confirm the source value and intended numeric value before calculating.
4. Rounded denominator: Keep the result approximate if you calculate it.
5. Wrong observation: Restore the intended source reference before recalculating.

A value type tells the sheet whether a cell contains a number, text, or another kind of value. A count can look numeric while the sheet stores it as text. Inspect its type and source rather than assuming that its visible appearance is enough.

A compact count label is also different from an exact source count. Do not silently turn a rounded public label into an exact integer. If the evidence supports only an approximate count, preserve that limit beside the result.

Next, inspect the stored result separately from its percentage display. A display format changes how a stored number appears. It does not establish that the stored number uses the intended units.

For an ordinary percentage display, store the ratio. The worked example below stores 0.1 and displays 10%. Multiplying the ratio by 100 first, then applying percentage formatting, applies the scaling twice.

Check the stored value directly. A screenshot of a formatted percentage does not show whether the formula uses the correct source cells or units.

## 6. Calculate the worked example by hand

Original fiction. All values and spreadsheet errors in this worked example are invented. They are not a product run or a real creator result.

The fictional sheet’s written rule is `(likes + comments + shares) / views`, displayed as a percentage.

| Count | Fictional source value | Required role |
| --- | --- | --- |
| Likes | 90 actions | Numerator component |
| Comments | 10 actions | Numerator component |
| Shares | 0 actions | Numerator component |
| Views | 1000 views | Denominator |
| Followers | 500 followers | Outside this formula |

First, calculate the numerator:

90 + 10 + 0 = 100 action counts.

Next, divide by the required denominator:

100 / 1000 = 0.1.

The stored ratio is 0.1 action counts per view. With percentage formatting, the sheet displays 10%. This does not mean that 10% of unique viewers engaged.

Now check the wrong denominator:

100 / 500 = 0.2, displayed as 20%.

That arithmetic uses followers. It fails this sheet’s written rule, which requires views. Correct the cell reference rather than accepting the percentage because it looks possible.

Next, check the display error. If the formula multiplies 0.1 by 100, it stores 10. If that cell also uses percentage formatting, it displays 1000%. Store 0.1 for a 10% percentage display.

Finally, check the incomplete row in the same fictional example. It has 40 likes, unknown comments, 0 shares, and 1000 views. The required numerator is incomplete. Treating unknown comments as zero would produce a displayed 4%, but that result would not satisfy the written rule. Mark the result unavailable.

The decision is therefore specific: correct the denominator reference, correct the stored ratio or its display, and preserve the incomplete row as unavailable. These changes verify the sheet’s arithmetic. They do not verify an audience-quality claim.

## 7. Check copied formulas and save the correction log

A correct first row does not establish that the copied formulas are correct. Inspect rows at the start and end of each formula range. Also inspect rows around inserted data, changed formulas, and missing-value cases.

For each checked row, confirm that the formula uses that row’s intended cells. Look for a reference that stays on another row when it should move. Also look for a reference that moves when the declared rule requires it to stay fixed.

Compare both the formula and its result. Matching displayed percentages can hide different source references.

Keep a correction log with these fields:

| Log field | What to save |
| --- | --- |
| Row and post identity | The affected spreadsheet row and source post |
| Old formula | The expression before correction |
| Old stored result | The numeric result before display formatting |
| Old display | What the reader saw |
| Failed rule | The exact requirement that the cell failed |
| Corrected rule or formula | The intended calculation and references |
| Evidence | The source fields, values, time, and precision |
| New result and state | The corrected value, unavailable state, or undefined state |
| Reviewer | The person who checked the correction |

Record why a value changed. A reviewer must be able to distinguish a formula correction from a fresh source observation.

Recalculate only eligible rows. Keep incomplete rows visible with their reason. Do not replace blanks with zero or remove affected posts without a note. Preserve approximate results as approximate after recalculation.

If the sheet combines posts or accounts, review that aggregation separately. Correct row formulas do not establish a correct combined summary.

## 8. Accept the sheet against clear checks

Accept the corrected sheet when these checks pass:

1. The intended numerator, denominator, and display are written down.
2. Checked source records match the intended posts and collection times.
3. Required counts retain their missing and precision states.
4. Hand calculations match the stored ratios and displayed percentages.
5. Missing required values remain unavailable.
6. Zero denominators remain undefined.
7. Copied formulas retain the correct source references.
8. The correction log explains each change.

State the scope of the review. If you checked selected rows by hand, name those rows. Do not describe that sample as a manual check of every result.

Keep the conclusion within the evidence. The [guide to TikTok fake-follower claims](/en/articles/tiktok-fake-follower-claims/) explains why a low public ratio does not prove fake followers. A corrected spreadsheet also does not predict sponsorship return.

## Next action: deliver the corrected sheet and its evidence

Give the client the corrected sheet, written formula, field-to-cell map, and correction log. State which rows you checked by hand. List any rows that remain unavailable, undefined, approximate, or uncertain.

If source evidence is missing, identify the needed record before collecting more data. Use the [TikTok Metrics Scraper](https://apify.com/chronometrica/tiktok-metrics-scraper) when a new public post observation fits that need. Keep the new collection time visible. Then check the returned source fields before using them in the sheet.

The next decision is whether the sheet now follows its declared rule. A matching percentage alone is not enough; the source cells, missing-data rules, stored units, and display must agree.

## Sources

- [Public question about follower-versus-view formulas — published May 12, 2025; retrieved October 25, 2026; one public question, no authenticated read or demand estimate](https://www.reddit.com/r/influencermarketing/comments/1kl020c/engagement_rate_help_brand_deal_advice_food/)
- [TikTok Metrics Scraper public default build 0.1.36 and supplied contract evidence — checked October 25, 2026; documented inputs and fields, no runtime test for this guide](https://api.apify.com/v2/acts/chronometrica~tiktok-metrics-scraper/builds/default)
- [TikTok Creator Analytics Task — checked October 25, 2026; public page presence only, saved settings not authenticated and Task not run](https://apify.com/chronometrica/tiktok-metrics-scraper/examples/tiktok-creator-analytics)