Brand response time on public social comments gets promised far more often than it gets measured. On 2026-08-29 this study read every audience comment on 80 Instagram posts from 40 large brand accounts and timed every reply those brands made. On 2026-08-31 all 40 of the same brands’ support pages and Instagram bios were read by hand, to record what each one tells the public about reply speed.
The question was narrow. How fast do brands actually reply in public, and what can a team commit to out loud without the commitment being a lie?
Anyone who has been asked to put a response time into a client agreement, a profile bio, a support page or a weekend rota has needed the second half of that question answered. What follows is the method first, because none of the numbers are readable without it.
What was measured: 40 Instagram accounts, 80 posts, 11,668 comments
The measured set is 40 Instagram brand accounts, 2 posts each, collected on 2026-08-29, carrying 11,668 audience comments and 680 brand replies. The accounts come from four sectors: airlines, banking and fintech, DTC retail, and food and beverage. Those four were chosen because response culture differs sharply between them.
The selection rule was frozen before collection began.
- The candidate pool started as a deliberately oversized list of well-known accounts in the four sectors.
- Every candidate handle was resolved against Instagram. Handles that did not resolve, were private, or had comments disabled were dropped, and the drops are named below.
- Follower counts were measured on the collection date rather than taken from a third-party ranking. The top 10 accounts by measured followers in each sector formed the final set of 40.
- Each account needed at least 8 posts in the 90 days before collection, so there was enough comment volume to produce a median.
- The set was frozen once collection began. No account was added or removed afterwards.
The reporting rules are as tight as the selection rule, and they are the reason some figures below are missing rather than estimated.
- Answer rate and time to first reply are reported separately. They are never blended into one score.
- Only a threaded reply from the brand counts as a reply.
- An unanswered comment is right-censored, never counted as zero. The brand may have replied after collection ended.
- Posts are restricted to an age of 48 hours to 30 days, so no brand is judged on a post nobody has seen yet. That window is sampling hygiene and nothing else.
- An answer rate is quotable at 20 comments or more. A median is quotable at 10 answered comments or more.
- Comment coverage is recorded for every post, because pagination does not always return a post’s full comment total.
The calibration that set those rules
Four calibration runs on 2026-08-28 set every rule above, and each one exists because a first attempt produced a number that looked clean and was wrong.
Instagram returned comments ranked rather than in time order on both collection dates. On one gymshark post of 80 comments, the answer rate read 13.3% over the top 15 comments, 6.7% over the top 30, 3.3% over the top 60 and 2.5% over all 80. A capped sample therefore overstates responsiveness by up to 5.3 times. Collection was run uncapped as a result, which is the single most expensive decision in the method and the reason the totals here are large.
Post age distorts the same way. One starbucks post 0.04 hours old carried 4 comments and no replies, which would have scored as total silence from an account that had simply posted minutes earlier. Hence the 48-hour floor.
Small samples distort hardest of all. One citibank post produced exactly one comment, that comment had a reply, and the post scored a perfect answer rate on a sample of one. The 20-comment and 10-answered floors come directly from that post, and they are applied below without exception.
Coverage is imperfect even uncapped. Pagination stops before the reported total on some posts: one citibank post returned 15 of a reported 23 comments, a coverage ratio of 0.652. Coverage was recorded per post rather than assumed.
What was dropped from the pool, and why
- nandos: the handle does not exist.
- figs, oura and wise: the guessed handle resolved to an unrelated private account. figs and oura were corrected to wearfigs and ouraring. wise was left out.
- allbirds: 0 posts in the 90 days before collection.
- qatarairways and klarna: 7 posts in 90 days, below the 8-post activity threshold.
This method suits a reader who has to defend a response-time number to a client or a manager, and who therefore needs to know what the number excludes. It does not suit a reader who wants one industry figure to quote without its conditions attached. This study does not supply one, and no such figure was computed to fill the request.
The answer rate: 680 of 11,668 comments got a reply
Across 40 brand accounts and 80 Instagram posts collected on 2026-08-29, brands replied to 680 of 11,668 audience comments. That is an answer rate of 5.8% for this set, on this surface, on that date.
Thirty-four of the 40 accounts cleared the 20-comment floor, so 34 answer rates are quotable.
| Sector | Account | Comments | Replies | Answer rate |
|---|---|---|---|---|
| Airlines | lufthansa | 181 | 108 | 59.7% |
| Airlines | airfrance | 88 | 28 | 31.8% |
| Airlines | turkishairlines | 89 | 16 | 18.0% |
| Airlines | etihad | 54 | 9 | 16.7% |
| Airlines | klm | 85 | 10 | 11.8% |
| Airlines | british_airways | 46 | 3 | 6.5% |
| Airlines | emirates | 298 | 12 | 4.0% |
| Airlines | singaporeair | 190 | 3 | 1.6% |
| Airlines | delta | 324 | 0 | 0.0% |
| Airlines | americanair | 205 | 0 | 0.0% |
| Banking and fintech | citibank | 31 | 6 | 19.4% |
| Banking and fintech | revolut | 85 | 14 | 16.5% |
| Banking and fintech | bankofamerica | 31 | 2 | 6.5% |
| Banking and fintech | hdfcbank | 38 | 2 | 5.3% |
| Banking and fintech | nubank | 1,454 | 55 | 3.8% |
| Banking and fintech | paypal | 35 | 0 | 0.0% |
| DTC retail | mejuri | 31 | 15 | 48.4% |
| DTC retail | glossier | 218 | 48 | 22.0% |
| DTC retail | ouraring | 40 | 8 | 20.0% |
| DTC retail | brooklinen | 38 | 6 | 15.8% |
| DTC retail | wearfigs | 60 | 7 | 11.7% |
| DTC retail | warbyparker | 43 | 4 | 9.3% |
| DTC retail | away | 91 | 5 | 5.5% |
| DTC retail | gymshark | 97 | 3 | 3.1% |
| Food and beverage | wendys | 74 | 11 | 14.9% |
| Food and beverage | dominos | 43 | 4 | 9.3% |
| Food and beverage | chipotle | 164 | 14 | 8.5% |
| Food and beverage | starbucks | 3,267 | 239 | 7.3% |
| Food and beverage | subway | 34 | 1 | 2.9% |
| Food and beverage | tacobell | 184 | 4 | 2.2% |
| Food and beverage | kfc | 88 | 1 | 1.1% |
| Food and beverage | dunkin | 3,391 | 34 | 1.0% |
| Food and beverage | burgerking | 106 | 1 | 0.9% |
| Food and beverage | mcdonalds | 423 | 0 | 0.0% |
Answer rate for 40 large Instagram brand accounts, 2 posts each, comments read uncapped on 2026-08-29. For teams planning the content that generates those comments, tools used to create Instagram posts can be part of the broader publishing workflow. Accounts below the 20-comment floor are excluded.
Six accounts did not clear the floor, so none of them has a quotable answer rate: chase with 17 comments, wellsfargo with 11, ruggable with 6, monzo with 3, everlane with 3, and icicibank with 2. Two of those six posted a higher raw rate than most of the table, and both are reported here as unquotable rather than quoted, because a rate built on 11 comments or on 3 is not a rate.
Four accounts answered nothing on enough volume to state it plainly: mcdonalds replied to 0 of 423 comments, delta to 0 of 324, americanair to 0 of 205, and paypal to 0 of 35. Three more accounts also answered nothing, and all three sit below the volume floor, so their zero is not quotable either.
At the other end, lufthansa replied to 108 of 181 comments, 59.7%, the highest quotable rate in the set. mejuri replied to 15 of 31, 48.4%.
The sector contrast is worth reading as counts rather than as averages. Airlines hold the highest quotable rate in the set and two of the largest zero-answer records at once: lufthansa at 59.7%, against delta’s 0 of 324 and americanair’s 0 of 205. Food and beverage holds the two largest comment volumes in the set, and both are answered thinly: dunkin replied to 34 of 3,391 comments, and starbucks to 239 of 3,267.
A reader who wants a sector average will not find one here. This study reports per-account results, and no sector mean or sector median was computed from them.
How long a reply takes when it comes
For the accounts that did reply, the median time to a first reply runs from 6.3 minutes to 851.0 minutes. Thirteen of the 40 accounts cleared the 10-answered floor, so exactly 13 medians are quotable.
| Account | Replies | Comments | Median minutes to first reply |
|---|---|---|---|
| lufthansa | 108 | 181 | 6.3 |
| revolut | 14 | 85 | 6.3 |
| turkishairlines | 16 | 89 | 7.1 |
| emirates | 12 | 298 | 54.0 |
| mejuri | 15 | 31 | 90.0 |
| chipotle | 14 | 164 | 97.0 |
| airfrance | 28 | 88 | 114.8 |
| starbucks | 239 | 3,267 | 155.6 |
| wendys | 11 | 74 | 275 |
| klm | 10 | 85 | 362.4 |
| nubank | 55 | 1,454 | 406.3 |
| glossier | 48 | 218 | 603.2 |
| dunkin | 34 | 3,391 | 851.0 |
Median minutes to a first threaded reply, for the 13 accounts in the set with at least 10 answered comments. Same 40-account set, same 80 posts, collected 2026-08-29.
Speed and coverage are separate findings, and one pairing in the table makes that unmissable. lufthansa and revolut share a median of 6.3 minutes. lufthansa answered 59.7% of its comments and revolut answered 16.5%. A fast median describes the comments a brand chose to answer. It says nothing about the ones it walked past.
The airlines make the same point in reverse. airfrance answered 28 of 88 comments, 31.8%, with a median of 114.8 minutes. turkishairlines answered 16 of 89, 18.0%, with a median of 7.1 minutes. The faster account answered fewer.
At the slow end, dunkin’s median of 851.0 minutes works out at 851.0 divided by 60, or 14.2 hours, on 34 replies across 3,391 comments.
Twenty-seven accounts had fewer than 10 answered comments, so none of them has a quotable median. Several of the slowest-looking numbers in the raw record belong to that group, which is exactly why the floor exists: gymshark answered 3 of 97 comments, kfc answered 1 of 88, and subway answered 1 of 34. One reply cannot carry a median, however slow that reply was, and none of those figures is quoted here.
One reporting rule bears restating at this point. Every unanswered comment in this study is right-censored rather than counted as a zero-speed reply, because collection ended at a moment the brands did not choose. A brand may have replied on day three.
A median is worth stating for an account with at least 10 answered comments in the window it measures. An account with fewer than that has no median yet, and the honest answer is to say so rather than to publish a substitute figure.
What the same brands publish as a promise
Across the same 40 brands, 3 publish a reply-speed commitment, 7 publish staffed hours only, and 30 publish neither. All 40 support pages and Instagram bios were read by hand on 2026-08-31. Some pages were reached only after searching for the correct address, because the obvious one returned a near-empty page.
The three verdicts mean three different things, and readers routinely collapse them into one.
- A response commitment states how fast the brand will reply. Three of 40 brands.
- Availability only states when a channel is staffed. That is not a reply promise. Seven of 40 brands.
- None found means neither statement was present on a page that was read. Thirty of 40 brands.
The three published reply-speed commitments, in full
ruggable, on https://ruggable.com/help/contact-us, read 2026-08-31:
> “Message our support team below and we’ll get back to you within 24 hours, often sooner!”
That promise covers the support team’s message form. It is not a promise about public comments. ruggable’s measured record is 2 replies across 6 comments, which is below the volume floor and supports no rate.
gymshark, on https://support.gymshark.com/en/articles/11186132-contact-us, read 2026-08-31:
> “#UK123456) The email address used to place the order A clear description of the issue Any relevant images (if needed) We aim to respond within 24 to 48 hours.”
The quote is recorded exactly as the page’s text reads when the bulleted list above it is flattened, and the operative sentence is the last one: “We aim to respond within 24 to 48 hours.” The promise covers email and the contact form. gymshark’s measured record on public comments is 3 replies across 97 comments, 3.1%.
everlane, on https://support.everlane.com/en_us/how-do-i-contact-everlane-support-BJLN3Hp0s, read 2026-08-31:
> “You can use our Contact Form for email support, and we’ll get back to you as quickly as possible (typically within a few hours).”
That is the fastest stated commitment in the set, and it covers the contact form. everlane’s measured record on public comments is 0 replies across 3 comments, which is below the volume floor and supports no rate.
The seven brands that publish staffed hours only
| Account | Comments | Replies | Answer rate |
|---|---|---|---|
| mejuri | 31 | 15 | 48.4% |
| ouraring | 40 | 8 | 20.0% |
| wearfigs | 60 | 7 | 11.7% |
| british_airways | 46 | 3 | 6.5% |
| emirates | 298 | 12 | 4.0% |
| tacobell | 184 | 4 | 2.2% |
| dunkin | 3,391 | 34 | 1.0% |
The seven accounts that publish staffed hours and no reply speed, with their measured records from the same 80 posts collected 2026-08-29.
One brand in the whole set states any time window on Instagram itself. hdfcbank’s bio reads: “Never share OTPs with anyone but share your banking queries with us on DM anytime between 7 AM – 11 PM.” That states when the account is staffed, not how fast it replies. No account in the set had Instagram’s business hours field filled in, and no profile link pointed at a support page.
A reader drafting a public promise can use this section as a record of what peers have actually committed to in writing. It is not a template. Three brands published three sentences on three pages, and those sentences are reproduced above rather than rewritten into wording anyone should copy.
The gap between promise and practice is mostly an absence
No brand in this set of 40 publishes a reply-speed promise for public comments. All three published commitments cover email and the contact form. The gap for public comments therefore cannot be calculated as measured time minus promised time, because the promised side is empty. A social listening platform like YouScan can help teams identify untagged mentions, sentiment shifts, and conversations beyond their own accounts.
What can be compared is thin, and it is worth being exact about how thin. gymshark’s answer rate on public comments is quotable at 3.1% of 97 comments, and gymshark’s median is not, because 3 replies cannot carry one. ruggable’s 6 comments and everlane’s 3 comments support neither a quotable rate nor a quotable median. The sample under each of the three promises is too small to test the promise, so none of the three was tested.
The second finding is the one most likely to matter to a team writing its own bio. Seven brands publish staffed hours, and readers hear a reply promise in them. Their measured answer rates run from dunkin at 1.0% of 3,391 comments to mejuri at 48.4% of 31. Stated availability predicted nothing about whether a comment got answered.
The third finding is the largest by count. Thirty of the 40 brands publish neither a reply speed nor staffed hours on a page that was read.
everlane is the case that shows the reporting rules doing their job. everlane publishes the fastest commitment in the set, “typically within a few hours”, and answered 0 of 3 measured comments, and 3 comments is below the volume floor and cannot support a rate at all. Both halves of that sentence are true and neither is a verdict on everlane.
Nothing here shows brands missing their own targets. It shows that almost no brand makes a target for this surface, and that the few who do made it about email.
What you can commit to, and what it costs in coverage
A response-time commitment holds up when it names four things: the surface it covers, the hours it covers, the metric it is measured in, and the unit that metric is reported in. Time to first reply, reported as a median, on public comments, during stated hours. A promise missing any of those four is not testable, and an untestable promise is the kind that gets quoted back during a bad week.
Two commitments get conflated constantly, and they have to be separated. How many comments get answered at all is one commitment. How fast the answered ones get answered is a second, unrelated one. Effective communication depends on understanding the difference. This study reports them separately and never blends them, and a public promise inherits the same rule. lufthansa and revolut, at an identical 6.3-minute median and answer rates of 59.7% and 16.5%, are the proof that one number cannot stand in for the other.
The arithmetic that sizes the promise
The measured unit here is comments per post, because that is what was collected: 2 posts per account, read uncapped. Turning comments per post into comments per week takes an assumption about posting frequency, and that assumption is not measured. It is written out below so it can be replaced with a real number.
Take the lufthansa shape first. lufthansa’s 2 posts carried 181 comments and 108 replies, which is 90.5 comments and 54 replies per post. Assume five posts a week, an assumption and not a measurement. The arithmetic gives 452.5 comments and 270 replies a week, delivered at a median first reply of 6.3 minutes. A commitment to a first reply inside one hour during staffed hours sits far outside that median, so lufthansa’s measured behavior would clear it comfortably.
Now take the dunkin shape. dunkin’s 2 posts carried 3,391 comments and 34 replies, which is 1,695.5 comments per post. At the same assumed five posts a week the arithmetic gives 8,477.5 comments a week. A commitment to reply to every comment therefore commits the team to 8,477.5 replies a week. Assume one minute per reply, a second assumption and a fast one, and that is 8,477.5 divided by 60, or 141.3 hours a week spent replying. Three full-time schedules at 40 hours each come to 120 hours. dunkin’s measured record is 34 replies at a median of 851.0 minutes.
The arithmetic resolves to one question. At the volume the account actually receives, how many replies does the stated response time require inside the stated hours, and does the team have those hours. Both worked cases above are descriptions of measured behavior, not judgments of either brand.
When a published response time is safe to make
- The team has measured its own answer rate and median over a fixed window on the same surface it is promising.
- The promised hours match a rota that exists, weekends and holidays included.
- The promise names the channel it covers.
- The team reports the median and the answered share separately, and keeps reporting them.
When it is not
- An account with fewer than 20 inbound comments in its window cannot state an answer rate, and one with fewer than 10 answered comments cannot state a median. Those floors were set for this study and they apply to the reader’s own account too.
- An account with no weekend or holiday coverage cannot publish a 24-hour promise. That promise fails every Friday evening.
- A team that cannot see its volume before it commits should publish staffed hours instead, and should expect readers to hear a reply promise anyway. Seven brands in this set are in that position, which is exactly the gap a dedicated customer support channel is meant to close.
- An account with inbound volume that looks like dunkin’s or starbucks’ cannot promise a reply to every comment at any staffing level a social team realistically has. The honest commitment there covers a defined subset of comments, and the copy names the subset.
- A promise written for email does not transfer to public comments. All three commitments in this set were email promises, and none of those three brands answered enough public comments to support a median.
A response-time promise that survives contact with the public names one surface, one set of hours, and a median measured on the account’s own last 30 days. Reply speed is one question inside the part of social media that is other people: comments, direct messages, creators and moderation. This study covers the first of those, on one platform, on one collection date.
Coverage arithmetic sizes a promise, not a team
The arithmetic above exists to size a commitment, and it stops there. Once the question becomes how many people the volume needs, how the rota is built, or what the load does to the person carrying it, the subject has changed from response time to workload and team structure. This publication treats that as its own subject elsewhere.
What the method cannot see
Four limits apply to every figure above, and each one makes a specific thing invisible.
- Only threaded replies count. A brand that answers by posting a new top-level comment reads as silent here.
- Direct message replies are invisible to this method. No figure above says anything about how fast these brands answer a DM.
- Deleted comments and deleted replies are gone. Neither appears in the record.
- Comment coverage is not complete even uncapped, so a post’s collected comments can fall short of its reported total. Coverage was recorded per post, and one citibank post came in at 0.652.
Why the account ranking was measured rather than cited
Published rankings of top brand accounts are the kind of unsourced table that gets cited in a circle until it reads as fact. This publication argues against that, so follower counts here were measured on the collection date and the top 10 per sector were derived from those measurements. Anyone who repeats the method on a stated date can rebuild the same set.
Common questions
How fast do brands actually reply to comments on Instagram?
Median time to a first reply ran from 6.3 minutes to 851.0 minutes across the 13 accounts in this set with at least 10 answered comments, measured on 80 posts collected 2026-08-29. There is no single figure for all brands, and the 13 medians above are per account for exactly that reason.
How many comments do brands answer at all?
Brands replied to 680 of 11,668 audience comments on public Instagram posts in this set, an answer rate of 5.8% across 40 accounts and 80 posts collected 2026-08-29. Four accounts with enough volume to report it answered nothing at all.
Do brands publish a response time?
Three of the 40 brands publish a reply-speed commitment, and all three commitments cover email or a contact form rather than public comments. Seven more publish staffed hours only, and 30 publish neither on a page that was read on 2026-08-31.
What response time can I promise a client?
You can promise a median time to first reply, on a named surface, inside named hours, once you have measured your own account over a fixed window. Report the answered share alongside it, because the two numbers move independently: two accounts in this set share a 6.3-minute median on answer rates of 59.7% and 16.5%. Until you have 20 inbound comments and 10 answered ones in that window, you have no rate and no median to promise.
Can I promise a 24-hour reply if nobody works weekends?
You cannot. A 24-hour promise made without weekend or holiday coverage fails every Friday evening, and it fails in public where the unanswered comment sits. Publish the hours you actually staff instead, and state the response time inside those hours.
Does this study include DMs?
No. This study measured threaded replies to public comments only, so direct messages are invisible to it. One brand’s bio in the set mentions DM hours, and that is a published statement by that brand rather than a measurement of it.
Can I use these numbers as an industry benchmark for my own account?
No. These figures cover one platform, 40 accounts, 80 posts and one collection date, with a 20-comment floor, a 10-answered floor and per-post coverage that is not complete. Measure your own answer rate and your own median over a fixed window, and use the method above rather than the numbers.
Source ledger
Every figure in the tables and the prose above comes from this study’s own collection: 40 Instagram brand accounts, 2 posts each, 11,668 audience comments read uncapped on 2026-08-29, with support pages and bios read by hand on 2026-08-31. The claims about how Instagram returned comments, and about pagination stopping short of a reported comment total, are stated as what this collection observed on those dates rather than as platform documentation.
| Source URL | What it supports | Date fetched | HTTP status observed |
|---|---|---|---|
| https://ruggable.com/help/contact-us | The ruggable reply-speed quote. The quoted sentence was present in the page returned on this date. | 2026-08-31 | 200 |
| https://support.gymshark.com/en/articles/11186132-contact-us | The gymshark reply-speed quote. The sentence “We aim to respond within 24 to 48 hours.” was present in the page returned on this date. | 2026-08-31 | 200 |
| https://support.everlane.com/en_us/how-do-i-contact-everlane-support-BJLN3Hp0s | The everlane reply-speed quote. The quoted sentence was present in the page returned on this date. | 2026-08-31 | 200 |
| https://blog.recurpost.com/category/community-creators/ | Linked archive on community management. | 2026-08-31 | 200 |
| https://blog.recurpost.com/ | Linked publication home. | 2026-08-31 | 200 |
No platform documentation is cited, because no claim above rests on one. No third-party statistic is cited, because none was used.
