Creator analytics · Verification guide
Why Instagram Engagement Rates Differ Between Tools
A creator’s media kit, a HypeAuditor report and a campaign dashboard can display different engagement rates. Before choosing a partnership, establish whether they measure the same interactions, content and audience reference.
Same creator.
Different percentage.
A traceable explanation.
Method: comparison of official product documentation, a dated Meta changelog and a reproducible illustrative example. Findings describe published methods; they are not results of a live audit of a creator account.
Start with the exact metric, not the provider name
Record the screen, metric label, selected dates and capture time for each number. “Modash engagement rate” is too vague: its Discovery and Campaign screens use different aggregation methods. A creator-level rate and a rate for one sponsored Reel also summarize different content.
Modash documents a median for Discovery and per-creator averages for Campaign ER. Campaign video ER uses actual video views as its denominator; the regular ER uses followers. These distinctions can explain disagreement within the same product.
Sources: Modash campaigns: aggregation, exclusions and refresh intervals.
The question to ask before rejecting a creator
“Can you share the exact metric label, formula, content list and reporting period behind this percentage?”
A screenshot showing only a large percentage leaves the comparison unresolved. Ask for enough context to reproduce or interpret it, without requesting account passwords.
What the official methods actually count
Sources: HypeAuditor: calculator formulas, eligible content and median · HypeAuditor: August 2026 methodology update · Modash Discovery: formula, pinned posts, hidden likes and updates · Modash public calculator: two-month content window · Minter.io: formulas and connection-dependent repost data, December 15, 2025 · Whatagraph: daily-average engagement rate, February 24, 2026.
These are documentation comparisons, not a ranking of measurement accuracy. A broader numerator does not automatically make one tool better for every task: authorized account reporting and public creator discovery have different data access.
Reconcile the numerator before changing the denominator
Illustrative single post: 900 likes and 100 comments on a 100,000-follower account produce a 0.9% likes-only rate or a 1% likes-plus-comments rate. With 20,000 views, the latter numerator produces a 5% view-based rate. The interactions did not increase when the denominator changed.
Do not label one percentage as the creator’s “real ER” until you define the question and the evidence available.
How to read HypeAuditor’s 2026 method change
HypeAuditor’s August 2026 digest documents a revised method that selects a relevant content group, uses views when enough suitable video content is available, and otherwise uses followers. It calculates individual content rates and takes their median, starting from the latest 30 content pieces.
Sources: HypeAuditor: August 2026 methodology update.
Its help page, dated July 10, 2026, already describes views-based ER and a fallback note when recent Reels are insufficient. August is therefore a documented explanation of the update, not proof of an exact launch day.
Sources: HypeAuditor help: unavailable Reels ER and private interactions.
Before comparing an old report with a new one, record the calculation basis and any insufficient-Reels message. The public pages reviewed do not state a precise minimum Reel count or fully specify all eligibility rules. We cannot reconstruct that selection from a single percentage.
If you need an initial public-data check, open the calculator and retain the displayed method alongside the result. For subscription access, pricing and operational fit, read our HypeAuditor buying assessment.
Match the publications, not just the date range
A month containing three Reels and a month containing thirty are different samples. Similarly, “latest 30 content pieces” is not equivalent to “all posts from the last two months.” Record the actual post links where available and distinguish Reels, static posts, Stories and carousels.
Pinned posts are not universally excluded
Modash states that pinned posts are included when they fall within the analyzed content window. Do not remove them from your reconstruction merely because they are pinned; follow the method you are checking.
Sources: Modash Discovery: formula, pinned posts, hidden likes and updates.
For your own campaign analysis, label sponsored and organic content separately when known. That is an editorial recommendation for clearer interpretation, not a claim that every provider automatically separates the two. A viral organic post may answer a different question from the sponsored placements you are considering.
Document exclusions in plain language: “three Reels lacked a usable view count” is more informative than silently reducing the sample from twelve posts to nine.
Three valid-looking summaries can give three different rates
The following invented example isolates aggregation. Each row counts the same interaction categories; all views use the same assumed platform scope and capture time. It is not a real creator, a market benchmark or a reproduction of a provider’s complete algorithm.
A median of ratios is not generally a ratio of medians. The pooled rate also gives publications with more views more influence. Publication A accounts for most views in this example and has the smallest rate, pulling the pooled result down.
The unweighted mean of the three post rates is 12%. It gives each publication the same weight, whereas the pooled rate weights them by views. None of these summaries identifies unique engaged people across the three posts.
Download the illustrative input data and formulas. In a spreadsheet, calculate each row’s ratio first and format it as a percentage. To obtain the median of post rates, apply MEDIAN to that ratio column. Do not multiply by 100 again if percentage formatting already performs the display conversion.
HypeAuditor’s digest describes individual-content ER followed by a median. This explains why dividing two separately displayed summary counts is not sufficient to reproduce its reported ER.
Treat unknown, excluded and zero as different states
A measured zero says the interaction count is zero. An unavailable value says the count is unknown. An excluded publication contributes nothing to that particular summary. Substituting zero for a hidden count changes the question and can depress your reconstruction artificially.
Modash states that hidden likes can prevent an accurate Discovery ER and that Campaign average ER excludes content with hidden likes, Stories and creators without posts. Record which of these situations applies before interpreting an absent percentage.
Sources: Modash Discovery: formula, pinned posts, hidden likes and updates · Modash campaigns: aggregation, exclusions and refresh intervals.
HypeAuditor’s analytical ER excludes saves and shares, which its help describes as private data only fully visible to the account owner. A creator-supplied report containing those interactions therefore cannot be compared directly with that public-data numerator.
Sources: HypeAuditor help: unavailable Reels ER and private interactions.
Evidence to request
Ask for the metric definition, post-level values where available, the number of included publications and any exclusion messages. Keep unavailable values blank and flag the reason. Request an authorized Insights export or a contextual screenshot when public data cannot resolve the question.
Check whether views include Facebook and paid distribution
A Reel shared to Facebook can create a platform-scope mismatch. Meta’s Instagram Platform changelog distinguishes facebook_views from crossposted_views: the former reports Facebook views; the latter combines Instagram and Facebook views for crossposted Reels.
The April 22, 2026 entry documents aggregated video metrics, total_views and total_views_count, covering surfaces including Facebook crossposting and boosted media. It also describes combined Instagram/Facebook view_count for crossposted video accessed through Business Discovery.
Sources: Meta: Instagram Platform changelog, including April 22, 2026.
This updates an older changelog statement that Business Discovery view counts remained Instagram-only. Preserve the date and API context when reading copied documentation; an earlier explanation can be accurate for its time and unsuitable as a current universal rule.
Modash’s campaign help attributes some in-app discrepancies to combined Facebook/Instagram views versus its Instagram-only count. That is Modash’s documented explanation of its integration, not a rule that every third-party tool excludes Facebook.
Sources: Modash campaigns: aggregation, exclusions and refresh intervals.
Practical check: ask which field the integration uses, whether the post was crossposted or boosted, and whether both interaction counts and views cover the same distribution. Adding Facebook views to the denominator while retaining Instagram-only interactions creates a different ratio.
The existence of a Meta field does not prove that a specific product uses it, has the required permissions, or backfills older data. Those implementation details require confirmation from the provider.
Record data freshness separately from capture time
Two reports captured today may contain counts last refreshed on different dates. Record both the capture timestamp and the update timestamp when available. If the latter is missing, mark it unknown instead of treating the report as live.
HypeAuditor’s help page dated October 16, 2025 lists ER, average likes and average comments among daily updates, while full report data are recalculated every thirty days. The documented cadence is not proof that every displayed field was refreshed at the moment of your visit.
Sources: HypeAuditor help: daily metrics and full report recalculation.
Modash Discovery help describes recalculation every two to four weeks. Campaign help describes approximately twelve-hour updates for recent content and notes that older content from prolific creators may stop refreshing. The screen therefore matters as much as the provider.
Sources: Modash Discovery: formula, pinned posts, hidden likes and updates · Modash campaigns: aggregation, exclusions and refresh intervals.
If a Reel is still accumulating views, compare counts captured at compatible times before treating a changed rate as a content-performance trend.
A verification worksheet for your next creator shortlist
Download the blank verification worksheet. Use one row per metric you want to compare; the template does not connect to an account or collect information.
Choose the appropriate outcome
- Compare: definitions, samples and capture conditions are aligned sufficiently for your stated task.
- Recalculate: source values are available, but the original summaries differ. Choose and label one common method.
- Request evidence: selection rules, counts or platform scope remain unknown. Keep the result unresolved rather than producing a false precision.
These outcomes are our proposed reporting workflow, not a vendor certification or an automatic fraud verdict. If an unexplained gap remains, send the provider the metric labels, post list, dates and calculations.
“We are comparing [metric] on [screen] with [other metric]. The attached posts cover [period], captured at [time]. We counted [interactions], divided by [denominator], and used [aggregation]. Please confirm content exclusions, data refresh and Facebook/paid coverage.”
What the reconciled rate can support
If the remaining question concerns audience authenticity, use our HypeAuditor checker interpretation guide. It separates AQS, Quality Audience and suspicious-account classifications from the ER calculation checked here.
A comparable engagement metric helps you interpret a shortlist. It does not establish audience relevance, sentiment, authenticity, sales or a fair fee by itself. Review the actual content and campaign objective alongside the numbers.
If your next task is deciding whether the platform suits your workflow, use our HypeAuditor review and limitations. If you are preparing an offer after assessing the creator, our creator sponsorship rate calculator addresses reach, format and commercial rights. An engagement percentage alone cannot set a final price.
Questions worth settling before you compare
Is a view-based rate always lower?
No. With an identical positive numerator, the larger denominator gives the lower rate. Views below follower count produce a higher view-based percentage; views above follower count produce a lower one.
Can I convert one account-level percentage into another?
Usually not from the percentage alone. A conversion requires the relevant source counts, selected content and aggregation. A single-post conversion does not reproduce an account-level median.
Does a 5% ER mean 5% of followers interacted?
Not automatically. The denominator might be views, interactions can include multiple actions by one person, and the people reacting are not necessarily followers. State the ratio rather than asserting a unique-person count.
Can a benchmark settle an unexplained difference?
No. First align the formula, content format, sample and relevant market. A benchmark calculated differently does not resolve why two reports disagree.
Sources
Official sources consulted October 19, 2026. Meta’s changelog was retrieved directly after the search reader returned a rate-limit error. Dates below describe source entries, not independently measured launch dates. Product integration behavior was not tested on a live account.
- HypeAuditor: calculator formulas, eligible content and median
- HypeAuditor: August 2026 methodology update
- HypeAuditor help: unavailable Reels ER and private interactions
- HypeAuditor help: daily metrics and full report recalculation
- Modash Discovery: formula, pinned posts, hidden likes and updates
- Modash public calculator: two-month content window
- Modash campaigns: aggregation, exclusions and refresh intervals
- Minter.io: formulas and connection-dependent repost data, December 15, 2025
- Whatagraph: daily-average engagement rate, February 24, 2026
- Meta: Instagram Platform changelog, including April 22, 2026
The source comparison and verification workflow are editorial analysis. The three-publication dataset is invented for transparent arithmetic and contains no creator data.