It's 11:40 at night. Someone is looking at a leather bag on a seller's page, and she types three words: “Is this available?”
The seller is asleep. Nothing wrong with that; she's a human being with a bedtime. At nine the next morning she opens the app, records a friendly voice note with the price, and waits. The message gets seen four hours later. Then nothing. No reply, no order, no goodbye.
Nobody writes that down as a lost sale. There's no failed checkout, no abandoned-cart email, no red number on a dashboard. The bag just stayed on the shelf.
I've watched this play out across a lot of seller inboxes, and the pattern is boring and brutal at once: the buyer didn't change her mind about the bag. She changed tabs.
The five-minute rule is really a five-minute window
The rule most people have heard came out of old lead-response research: reach a new lead inside five minutes instead of thirty, and your odds of actually connecting with that person and finding out whether they'll buy go up by something like twenty-one times. Same lead. Same offer. Same script. Only the clock moved.
Selling in DMs is harsher than that, because your buyer already has the phone in her hand. She messaged you from a story, mid-scroll, probably from bed. Her interest has a shelf life measured in minutes, and there are four other sellers one search away with a nearly identical product and a working thumb.
So five minutes isn't an average you aim for. It's a window that either catches her while she still wants the thing, or closes. Averages hide the damage beautifully: answer half your DMs in one minute and half in nine hours, and your reported average looks respectable while half your revenue quietly walks out.
Put a number on your own slow replies
Most sellers argue about response time in feelings. Do it in arithmetic instead. It takes two minutes.
- Count last week's DMs that asked about price, size, or availability. Say 120.
- Count how many you answered inside five minutes. Say 30.
- Of the 90 that waited, how many turned into an order? If it's 9, that's 10%.
- Of the 30 fast ones, how many turned into an order? If it's 12, that's 40%.
The gap between those two rates, multiplied by your average order value, is your weekly cost of being slow. With a $60 basket, those 90 slow conversations should have produced something like 36 orders instead of 9. Call it $1,600 a week, gone, with no notification anywhere to tell you about it.
Run it on your own inbox before you argue with me. I have yet to see a seller do this exercise and come back saying speed didn't matter. Usually they come back annoyed, which is the correct reaction.
Buyers don't wait nearly as long as you think
Sellers tend to picture a patient customer who will check back tomorrow. That customer is rare. When we looked at how long DM customers actually wait before giving up, the honest answer was: not long, and far less than the seller's own guess. People send the same question to two or three shops and buy from whoever answers first with a straight price.
A second problem hides underneath. The hours that produce the most DMs are usually not the hours you're working. Late evening is a peak — that's when someone finally sits down, scrolls, and decides she wants the bag. If your inbox is dark from 10pm to 9am, you're absent for the exact stretch when buying intent is at its highest.
There's a psychological layer here that sellers underrate. A reply that lands in ten seconds says something about the shop behind it: this person is organized, the product probably exists, the parcel will probably arrive. A reply that lands tomorrow afternoon says the opposite, before you've made a single argument about quality or craftsmanship.
That's why fast answers lift conversion even when the price is identical to the competitor's. You're not only winning a race against another shop. You're removing the buyer's quiet worry that she's about to send money to a stranger who might vanish.
The three questions that decide the sale
Most DM conversations hang on the same three things: is it available, what does it cost, and how fast can I get it. Leave any one of them unanswered and the thread dies politely: a “thanks” with no order, or silence. We wrote about the three unanswered questions that quietly kill sales, and the fix is unglamorous. Answer all three in your first reply, before she asks.
Speed without substance doesn't rescue you either. A two-second “Hi dear, DM sent 🙏” is fast and worthless. It burns the window without moving the buyer one step closer to paying.
What a good fast reply contains
The reply that converts is short, specific, and closes the loop. It looks like this:
- Yes or no on stock, for the exact variant she asked about, in the first line.
- The price, written as a number, not “price in DM.”
- Shipping time and cost, so she doesn't have to ask a second question.
- One question back — colour, size, city — that pushes the conversation toward an order.
Four lines, sent in under a minute. That plain format beats most of the elaborate sales scripts I've watched sellers try to memorize.
“I'll answer as soon as I can” stops working at some point
Here's the trap. A seller with 30 DMs a day can be fast by being disciplined. At 100 a day, discipline runs out. You're photographing products, packing orders, arguing with a courier, and the inbox turns from a conversation into a queue you triage.
The usual fixes each break somewhere. Hiring someone helps until 6pm and costs a salary every month. Saved replies are fast but rigid: they don't know your stock, so they can't say “yes, the tan one is here.” Keyword bots answer the wrong question with total confidence, which is worse than silence, because now the buyer thinks you're incompetent rather than busy.
What holds up is an assistant that knows your catalogue and your prices, replies in seconds at 2am in your own tone, and hands the thread to you the second it touches something it shouldn't decide: a discount request, a complaint, a custom order. We compared AI and human reply speed on real seller inboxes, and the distance was not subtle. Seconds against tens of minutes.
Where automation should stop
I won't pretend automation is risk-free. The failure I see most often is a seller who switches everything on, never reads the transcripts, and discovers two weeks later that the assistant has been quoting last season's price to everyone who asked.
So draw the line on purpose:
- Automate the boring 80%: availability, price, sizes, shipping, store address, payment methods, order status.
- Escalate the money-sensitive 20%: discounts, returns, angry customers, anything custom.
- Read ten real conversations a week. Not the metrics, the actual text. You'll find wrong answers faster than any report will show you.
Do that, and speed stops being a personality trait you have to maintain at midnight. It becomes a property of your shop.
Try it for one week
Pick the seven questions you get asked most. Write the exact answer to each one, with real numbers in it. Turn on instant replies for those seven only, and only during the hours you're offline. Everything else stays manual.
Then compare two weeks of orders at a similar DM volume. If nothing moves, you've lost a week and learned something. If your inbox behaves like most of the ones I've seen, it's the night-time conversations that change first, and the change is not small.
If you'd rather not build that yourself, Vardast connects to your Instagram, WhatsApp, Telegram, and website, learns your products and prices, and answers customers in your own voice while you sleep — then steps aside the moment a real person is needed.