Search for how quickly you should respond to a lead and you will find the same statistics repeated everywhere: respond within five minutes and you are 100 times more likely to connect, 21 times more likely to qualify the lead.
Those numbers come from research published in 2011, based on outbound phone calls to web form submissions. They were good research. They are also fifteen years old, they predate the shift of first contact to messaging, and nobody has replaced them for the channels people actually use now.
So we measured it.
What we measured
We analysed 43,971 Instagram DM conversations that began with an inbound message from a lead between January 2025 and September 2026, across hundreds of agencies and businesses using Inflowave.
For each conversation we measured the gap between the lead's first message and the first reply they received, then tracked what happened afterwards.
This is Instagram direct messages specifically. It is not phone, not email, not web form follow-up. Where the old studies measured whether a rep managed to make contact, we are measuring whether the conversation continued.
Finding 1: the decay is real, but far gentler than the phone-era numbers
| Time to first reply | Conversations | Lead replied again |
|---|---|---|
| Under 1 minute | 18,283 | 89.6% |
| 1 to 5 minutes | 3,386 | 87.8% |
| 5 to 15 minutes | 2,441 | 83.1% |
| 15 to 60 minutes | 3,798 | 79.4% |
| 1 to 4 hours | 4,493 | 81.9% |
| 4 to 24 hours | 5,339 | 76.3% |
| 1 to 3 days | 2,634 | 71.9% |
| Over 3 days | 3,597 | 72.9% |
The gap between replying inside a minute and replying after a day is 17.7 percentage points. That is a real, consistent decline and it is worth money.
But it is nothing like a hundredfold collapse. Wait three days to answer an Instagram DM and roughly seven in ten leads still reply. On the phone-era model those conversations should be effectively dead. They are not.
This makes sense once you think about the medium. A missed phone call is gone. A DM sits in an inbox with a notification badge, and the lead can pick it up whenever they choose. Messaging is asynchronous by design, and it forgives lateness in a way a ringing phone never did.
If you have been told your leads die in five minutes, that is not true on Instagram. The urgency is real but the cliff is a myth.
Finding 2: speed no longer buys you contact, it buys you conversation
This is the part we did not expect.
Replying quickly barely changes whether someone answers. It dramatically changes how much they say.
We compared conversation depth in the seven days after our first reply, using a fixed observation window so that older and newer conversations get exactly the same amount of time to develop. Without that control the 2026 figures look far more dramatic than they are, because recent conversations have had less time to accumulate messages.
| 2025 | 2026 | |
|---|---|---|
| Conversations analysed | 15,237 | 27,565 |
| Messages from lead, replied in under 1 min | 13.83 | 18.96 |
| Messages from lead, replied after over 1 day | 10.95 | 4.52 |
| Ratio | 1.26x | 4.19x |
| Lead replied at all, fast | 91.5% | 87.2% |
| Lead replied at all, slow | 70.8% | 66.3% |
Read the ratio row.
In 2025 a slow reply cost you a slightly shorter conversation. About 11 messages instead of about 14. Annoying, not fatal.
In 2026 a slow reply costs you roughly three quarters of the conversation. Under 5 messages instead of nearly 19.
The reply-rate gap held steady across both years at about 21 points. The depth gap more than tripled. Whatever changed, it did not change whether people answer. It changed how long they stay.
What we think is happening, and what we cannot prove
We can measure the effect. We can only reason about the cause, so treat the next paragraph as interpretation rather than data.
The likely explanation is that expectations moved. Enough businesses now answer instantly, including automatically, that a slow reply no longer reads as normal. The lead has already messaged three competitors, two answered immediately, and by the time the third replies the decision is further along. They still answer out of politeness. They just do not invest in the conversation.
If that is right, the practical meaning is uncomfortable: the value of speed is inflating, and it will keep inflating as automated first responses spread. The advantage of being fast is not that you catch people. It is that you are still in the running when they decide.
Four things that would make us wrong
We would rather flag these than have someone else find them.
This measures conversation, not revenue. We tracked whether leads kept talking. We did not tie these conversations to closed deals, because that data is not reliably linked across the businesses in the sample. A longer conversation is a reasonable proxy for a warmer lead. It is not a sale.
Survivorship. Every conversation here is one where somebody eventually replied. Leads who were never answered at all are not in the population, so the true cost of slowness is probably understated rather than overstated.
The fast group is substantially automated. 41.6% of first replies land in under a minute, which is not a human typing speed at that volume. So the honest reading of the top row is closer to "automated speed versus human speed" than "quick people versus slow people". That does not invalidate it, because the lead cannot tell the difference and is responding to elapsed time either way. But it is not a study of well-organised humans.
2026 is a partial year. It runs to September. We excluded conversations starting after 25 August so that every conversation had at least seven days to develop, but a partial year is still a partial year.
What to do with this
If you take three things:
Stop repeating the five-minute cliff for messaging. It is a phone statistic from 2011 and your own inbox disproves it. Leads on Instagram stay reachable for days.
Speed is now a depth lever, not a contact lever. If you are optimising for whether someone replies, you are optimising the thing that barely moved. The prize is the conversation that runs to 19 messages instead of 4.
The gap is widening, so this is worth acting on now. A 1.26x penalty became a 4.19x penalty in a single year.
Methodology
Population: conversations whose first message was inbound from a lead, excluding those flagged as spam, where at least one outbound reply followed. Response time is the interval between that first inbound message and the first outbound reply. Finding 1 covers first messages from 1 January 2025 onward. Finding 2 caps first messages at 25 August 2026 and counts lead messages only within seven days of our first reply, so both year cohorts have identical observation windows. Channel is Instagram direct messages; no other channel had sufficient volume in the period to report separately. Figures were computed on 12 September 2026.
Reuse. These figures are free to cite with attribution to Inflowave and a link to this page. If you want a cut we have not published, such as a specific vertical or a different time window, ask and we will run it.
Inflowave sells a CRM, so we are an interested party. We have published the method, the sample size, the confounds and the numbers that undercut our own commercial interest, specifically that leads do not die nearly as fast as the industry claims. Check the workings and disagree with them if they do not hold up.
