Decide what question the archive should answer
Before requesting an account export, write down the question you are trying to investigate. You might want a record of your own posts, a timeline of account changes, or a way to organize information you have already supplied to a platform. Do not assume an export can answer every question about audience behavior or reveal information the platform does not provide.
This guide explains a cautious reading workflow. Frenzlab does not accept, upload, or parse account archives. The engagement rate calculator accepts manually entered numbers only. Keep a private export on your own device and avoid sending the entire archive to an unfamiliar service just to obtain a simple summary.
The categories, formats, date ranges, and availability of data must be verified in the current official export instructions. [VERIFY: Instagram Help Center — accessing and downloading your information — https://help.instagram.com/; TikTok Support — requesting your data — https://support.tiktok.com/] A familiar filename in someone else's archive is not evidence that your current download will contain the same records.
Request the data through an official route
Start with the platform's official help center or verified account settings. Read the current instructions for requesting information and record the choices offered, including the selected account, date range, categories, and output format where available. [VERIFY: Instagram Help Center — download your information through Accounts Center — https://help.instagram.com/; TikTok Support — download your data — https://support.tiktok.com/]
Do not enter credentials into a third-party export helper merely because it promises a faster archive. This site does not provide such a helper. If you manage a client's account, obtain the owner's authorization and agree how sensitive material will be handled before requesting more data than you need.
Keep a request note with the date, intended scope, and any official confirmation. Availability and timing should be taken from the current official process, not guessed from a blog. [VERIFY: Instagram Help Center — information download availability and timing — https://help.instagram.com/; TikTok Support — data request preparation and availability — https://support.tiktok.com/] An unresolved request is not a reason to invent missing data.
Preserve the original and work on a copy
Store the downloaded archive in a private location and keep an unchanged original. Make a separate working copy before reorganizing files or opening them in tools that may save changes. Give the folder a descriptive date and account label that does not expose sensitive identifiers unnecessarily.
For example, a private folder named account-export-2026-09-review can contain the untouched archive, a working copy, and your notes. If you later discover a parsing error, the original lets you repeat the analysis. Without it, you may not know whether a strange value came from the source or from your own editing.
Treat unexpected executable files, external links, or requests to install software with caution. You do not need to run arbitrary code simply to read a text-based record. If a file is unfamiliar, inspect its type and consult the official export documentation before opening it in a way that executes content. This is a precautionary workflow, not a claim that every archive has the same structure.
Understand the formats you actually receive
An HTML export is intended to be read as linked pages, while JSON represents structured data and CSV represents rows and fields. The precise schemas and content are defined by the export you received, not by this guide. General format behavior can be checked in the specifications. [VERIFY: IETF — RFC 8259, JSON data interchange format — https://www.rfc-editor.org/rfc/rfc8259; IETF — RFC 4180, CSV format — https://www.rfc-editor.org/rfc/rfc4180]
Do not assume a spreadsheet can open every file correctly just because it contains readable text. A JSON object can contain nested arrays, and a CSV field can contain commas or line breaks inside quoted text. Use a reader suited to the format and inspect a small sample before processing a whole archive.
| Observed format | Useful first action | Avoid |
|---|---|---|
| HTML pages | Read the local index and a few linked records | Treating every link as safe to follow externally |
| JSON | Inspect keys, nesting, and a small sample | Flattening nested records without understanding them |
| CSV | Confirm delimiter, encoding, and column meanings | Assuming a comma always separates two fields |
| Media folders | Check filenames against associated records | Inferring dates solely from a file's display order |
These are starting actions, not a promise that the platform offers every format listed.
Read dates and time zones explicitly
Look for the representation used by the actual record: a formatted date, a timestamp, or a field with an explicit offset. Do not silently interpret every timestamp as your local time. Standard timestamp conventions need careful handling. [VERIFY: IETF — RFC 3339, date and time on the Internet — https://www.rfc-editor.org/rfc/rfc3339]
Imagine two records describe the same instant, one as 09:00 at UTC and another as 11:00 with a positive two-hour offset. The displayed clock times differ while the instant can be the same. If you sort only by the visible hour, you may invent a sequence that the records do not support.
Also separate a content creation time from a download time, update time, or file-copy time. A file's local modification date may reflect your own handling rather than the event you want to study. Determine which field actually represents the event before building a timeline. If the archive does not explain a field, preserve the uncertainty in your notes rather than choosing the most convenient meaning.
Distinguish identifiers from display names
An identifier and a display label serve different purposes in a dataset. Do not merge records solely because the visible names look similar, or assume a changed display name necessarily means a different entity. The export's actual field definitions need verification. [VERIFY: Instagram Help Center — downloaded information categories and records — https://help.instagram.com/; TikTok Support — downloaded data categories — https://support.tiktok.com/]
Create a small mapping table only when you have evidence for the relationship. Keep the original value alongside any human-readable label you add. If two records are ambiguous, mark them as unresolved rather than combining them to make the spreadsheet cleaner.
For a client report, remove private identifiers that are unnecessary to the question. You can often summarize a pattern without sharing message contents, contact details, or a full archive. Decide what the recipient needs to know and share the minimum relevant extract, subject to the owner's authorization.
Check coverage before calculating totals
Confirm the requested date range, included categories, and whether the archive describes a snapshot or a specific interval. Completeness and recent-data availability must be checked in the platform's documentation. [VERIFY: Instagram Help Center — information download scope — https://help.instagram.com/; TikTok Support — download your data, availability limitations — https://support.tiktok.com/]
Suppose you count twelve post records in a selected period but remember publishing fifteen items across several surfaces. The difference does not automatically prove deletion or missing evidence. First check whether the surfaces, dates, and categories match the request. Write twelve records found in this selected export category until you can support a broader claim.
If you combine two exports, look for overlap before adding their totals. Two overlapping date ranges can count the same record twice. Deduplicate only using a rule you can explain and keep the original rows so the operation can be reviewed. A clean total should come from a documented method, not from deleting inconvenient duplicates by eye.
Work through a modest numerical example
Imagine a manually verified table contains five comparable posts with interaction totals of 40, 60, 80, 100, and 120. Their sum is 400 and their average is 80. If you enter those interactions against a combined denominator, label the resulting rate as a pooled calculation. If you average five individual rates instead, state that method separately.
Now suppose one row belongs outside the requested period. Removing its 120 interactions leaves 280 across four posts, averaging 70. The correction changes both the numerator and sample size. Record why the row was excluded rather than silently replacing the earlier result. That note is part of the analysis, not clutter.
Use the reach-drop guide when comparing periods. An export can organize the records you have, but it does not by itself establish why a distribution change occurred. Keep measurement and explanation as separate stages.
Do not infer private audience behavior without evidence
An archive should not be marketed as a way to discover hidden profile viewers, private viewers, or other information not actually provided. Whether a platform supplies a particular category must be checked in its official documentation. [VERIFY: Instagram Help Center — information available in account downloads — https://help.instagram.com/; TikTok Support — data download categories and profile-view features — https://support.tiktok.com/]
A name appearing in one record does not establish every action that person took. Read the record's meaning before drawing a conclusion about someone else. Avoid turning an ambiguous log into a personal accusation or a public list of supposed viewers.
When the analysis is finished, keep only the working copies and extracts you need, according to your own privacy and retention responsibilities. Share conclusions with their scope, method, and limitations attached. Our editorial policy follows the same principle: a readable table should make uncertainty easier to see, not conceal it behind an authoritative-looking number.
Make the analysis reproducible
Keep a short method note next to your results. Name the source export, the selected category, the date interpretation, the inclusion rule, and any rows you excluded. If you used a spreadsheet formula, retain it rather than replacing every cell with its displayed result. A colleague should be able to understand how the summary was produced without receiving unrelated private material.
For example, if you excluded two records because they were outside the selected reporting period, state that reason and retain their references privately. If a later export changes the result, repeat the same method before deciding that the underlying activity changed. Reproducibility is an editorial discipline, not a guarantee that the source archive is complete.