Compare Instagram Followers vs Following: Sets, Ratios, and Formulas
Model Instagram Followers and Following as two sets, calculate mutuals and one-way relationships, and verify the results in Google Sheets.
On This Page
Instagram Followers and Following point in opposite directions. Followers are accounts that follow the profile; Following is the list of accounts the profile follows. The relationship is not automatically mutual.
Let:
F= the unique accounts observed in the current Followers file;G= the unique accounts observed in the current Following file.
Then the complete current relationship model is:
Mutuals = F ∩ G
Follower-only = F − G
Following-only = G − F
All distinct relationship accounts = F ∪ G
All one-way relationships = (F − G) ∪ (G − F)This page is a formula and verification reference. If you only want the step-by-step list of accounts that do not follow you back, use Who Doesn't Follow Me Back on Instagram.
The four relationship buckets
Following-only is a current relationship, not proof of a historical unfollow. A historical change requires an older and newer Followers snapshot.
Why two totals are not enough
Knowing only |F| and |G| cannot reveal the mutual count. Two profiles can both show 1,000 Followers and 800 Following while having very different intersections.
The total gap is useful arithmetic, but it does not identify a relationship bucket:
Balance gap = |F| − |G|Do not call that gap the number of non-followers.
Count identities that should always balance
Define:
M = |F ∩ G|— mutual count;FO = |F − G|— follower-only count;GO = |G − F|— following-only count;U = |F ∪ G|— union count.
The result should satisfy:
|F| = M + FO
|G| = M + GO
U = M + FO + GO
One-way count = FO + GOIf these do not balance, check duplicates, blanks, mixed identity keys, and incorrect range references.
Five useful ratios
Here, card(F) and card(G) mean the number of unique accounts in sets F and G.
When a denominator is zero, mark the ratio as not applicable. Do not report infinity or manufacture a zero percentage.
Worked example
These IDs are fictional:
F = {101, 102, 103, 104}
G = {102, 103, 105}The results are:
- Mutuals:
{102, 103};M = 2 - Follower-only:
{101, 104};FO = 2 - Following-only:
{105};GO = 1 - Union:
U = 5 - Followers/Following ratio:
4/3 = 1.33 - Follow-back coverage:
2/3 = 66.7% - Reciprocal share:
2/4 = 50% - Jaccard overlap:
2/5 = 40%
Those values describe only the fictional sets. They are not Instagram benchmarks or product test results.
Google Sheets formulas
Assume Followers!A2:A and Following!A2:A contain deduplicated account IDs. Google documents COUNTIF as a conditional count and FILTER as returning rows that satisfy the supplied conditions.
Mutuals
=FILTER(Followers!A2:A, Followers!A2:A<>"", COUNTIF(Following!A2:A, Followers!A2:A)>0)Follower-only
=FILTER(Followers!A2:A, Followers!A2:A<>"", COUNTIF(Following!A2:A, Followers!A2:A)=0)Following-only
=FILTER(Following!A2:A, Following!A2:A<>"", COUNTIF(Followers!A2:A, Following!A2:A)=0)COUNTIF is not case-sensitive. That helps if username capitalization varies, but it does not solve renames, deletions, deactivations, or a poor normalization rule. FILTER returns #N/A when no row matches; you can wrap a valid empty result in IFNA, but first make sure the error is not caused by misaligned ranges.
Build comparable source sets
Instagram Follower Exporter runs one audience direction at a time. To build F and G:
- Use the same source profile and a similar collection window.
- Export Followers and Following in separate runs.
- Keep the same 100, 200, 500, or 1,000-row ceiling.
- Keep the same default account-ID deduplication behavior.
- Record each task's terminal state and confirmed rows.
The 14 output fields are id, pk, username, profileUrl, fullName, profilePicUrl, isVerified, followedByViewer, requestedByViewer, audience, sourceProfile, sourceProfileUrl, collectedAt, and rowIndex.
Use id as the comparison key. Use a normalized username only as a fallback. Do not compare fullName, profile-image URLs, or rowIndex: display names can repeat, images change, and row numbers are only file positions.
Verification checklist
- Both files describe the same
sourceProfileandsourceProfileUrl. - One file is Followers and the other is Following.
- Ceilings and deduplication settings match.
- Collection times are close enough for the intended question.
- Terminal state and confirmed rows are recorded;
partialis not rewritten as complete. - All count identities balance.
- Important rows in every bucket are manually sampled.
- Original files remain unchanged; formulas live in a working copy.
Limits
Web Tidy confirms at most 1,000 rows per run. For a larger account, the ratios describe two bounded observed sets, not the complete social graph. Login, challenge, rate limit, private/unavailable targets, and lost context stop collection; transient errors and HTTP 5xx responses are retried at most three times.
A completed selected task does not guarantee coverage of the profile's displayed total. Current sets do not contain historical event time or reasons. None of these ratios proves engagement, reach, account quality, revenue, or permission to contact people.
The product does not output JSON, monitor on a schedule, enrich email/phone data, sync a CRM, automate follows/unfollows/messages, or bypass private profiles and access controls.
Frequently asked questions
Does more Followers than Following mean more mutuals?
No. Mutuals are the intersection |F ∩ G|, which cannot be derived from two totals alone.
What is the Followers-to-Following ratio?
|F| / |G|. It compares total sizes only. It is not an official Instagram health score.
Which ratio best describes reciprocity?
Use M/|G| for the share of your Following that follows back, M/|F| for the share of Followers you follow, or M/U for overlap relative to all distinct accounts. State the denominator.
Can I compare usernames instead of IDs?
Only as a fallback. Account IDs are more stable; usernames can change.
Can these formulas find recent unfollowers?
No. Compare older and newer Followers files for that separate question.