The Human Algorithm.

India Reply Report 2026We asked 83 Indian brands one simple question.

Is it in stock? Is COD available? Here is who answered, how fast, and what it means for your own inbox.

Madhur Tyagi
Madhur Tyagi
1 November 2026 · 9 min read
Download the PDF
HelpTuesday
Hi, is this kurta in stock in M? And is COD available to Kanpur?
10:42 PM · Sent
next morning
No reply yet
Neha is a composite of many buyers. The numbers below are not.

Quick answer

In October 2026 we asked 83 Indian D2C brands one pre purchase question: is this product in stock, and do you offer COD? 51 responded (61.4%) and 32 never replied (38.6%). 20 fully answered (24.1%). The median human reply took 235.5 minutes, about 4 hours, and 8 of 36 human replies took more than a day.

Written by Madhur Tyagi, founder of The Human Algorithm. 8+ years, 25+ brands, since 2017.

It is 10:42 on a Tuesday night in Kanpur.

Neha has had the same kurta open for two days. Her cousin’s wedding is in ten days.

She has two small questions. Is it in stock in M? Will they take cash on delivery to her pincode? She types them into the brand’s help box, presses send and puts her phone face down.

In the morning there is nothing. By lunch she has bought a similar kurta from a brand that answered. The brand she wanted never knew she was there.

We wanted to know how often Neha gets her answer. So for one week in October 2026, we asked 83 Indian D2C brands the same kind of question she did, and we timed every reply.

Neha is a composite. The question, the brands and every number in this report are real. No brand is named.

The question we asked

One message per brand, about a real product on its own site, checked as available right before we sent it. Most messages read like this:

Our question, word for word

Hi, is the [product] in stock, and do you offer COD to Delhi? Thanks

It is the most ordinary question in Indian ecommerce. Any team can answer it in under a minute, if someone reads it.

How we did it

  • 83 brands, one week. Sent between 25 and 31 October 2026 (no sends on 28 October). Fashion 18, beauty 18, food 16, jewellery 15, wellness 11, home 5.
  • One question per brand. Polite, no links, no pitch, no mention of a study. Signed by an assistant, from a normal email address.
  • The brand’s own channel. Site chat where it opened, else the support email printed on the brand’s own site. 76 by email, 7 by site chat. Contact forms all hit a CAPTCHA or gave no sign the message went through, so none counted.
  • No shortcuts. We never logged in, bought anything, gave a phone number or solved a CAPTCHA. Brands we could not reach that way were left out and noted as walls.
  • Replies counted to 1 November, 11:17 AM IST. About a week of waiting for the earliest brands, a full day for the last batch.
  • Every reply read one by one. A reply counts as a person when it names an agent or clearly speaks to our exact product in personal words. Chat tools that said they were AI, or answered in seconds with a set format, count as bots. Ticket mails, surveys and “ticket closed” notices count apart from answers. Templates that could be either are marked unclear.

What this is not. It is a small sample with one question each, mostly well known brands, in a busy festive month. It says nothing direct about WhatsApp speed. Read it as an honest signal, not a census.

The numbers

Base is 83 brands unless a card says otherwise.

61.4%responded at all51 of 83
38.6%never replied32 of 83
24.1%fully answered the question20 of 83
~4 hmedian human reply235.5 min, 36 replies
8 of 36human replies took over a day22.2%
7.2%asked for a phone number or order id6 of 83

What happened to 83 questions

Each brand counted once, by the most useful thing it sent back.

  • Fully answered20 (24.1%)
  • Partly answered16 (19.3%)
  • Replied, but no answer15 (18.1%)
  • Never replied32 (38.6%)

How long people took to reply

36 human replies, time from our message to the opening human reply.

Under 1 hour
8
1 to 4 hours
11
4 to 24 hours
9
Over a day
8

Who answered

51 brands that responded.

  • A person36 (70.6%)
  • A bot10 (19.6%)
  • Template, unclear3 (5.9%)
  • Ticket mail alone2 (3.9%)

Email or site chat

Share of each channel that got any response.

Email
57.9%
Site chat
100%

Email: 76 sent, 44 responded, all 36 human answers. Site chat: 7 sent, 7 responded, 6 bots and 1 unclear.

By category

Small groups. A hint, not a verdict.

CategoryAskedRespondedFull answer
Fashion18136
Beauty18135
Food16124
Jewellery1553
Wellness1160
Home522

5 things we found

Finding 1

Almost 4 in 10 brands never replied.

32 of 83 brands sent nothing back. Not a ticket, not a bot, not a “we will check”. We counted replies until 1 November, so the brands we wrote to on 25 October had about a week. The last batch had a full day.

Every one of those 32 questions came from someone holding a product page open. None of them shows up in a sales report. A lost order leaves no trace.

What this means for youCount the questions you did not answer, every week, next to the orders you did get. That number is your hidden leak.

Finding 2

A reply is not an answer.

Our question had two parts: is it in stock, and is COD available. 20 of 83 brands answered both (24.1%). 16 answered one part. The most common half answer on COD was “check your pincode at checkout”. That is homework, not help.

Another 15 brands responded without answering at all: 11 real replies that missed the question, and 4 ticket or template mails that did not answer it. Of 51 responses, 21 carried a product link or price.

What this means for youIf a buyer asks two things, give two clear answers in one message, and add the link so the next tap is the buy button.
A reply is not an answer.

Finding 3

The typical human reply took about 4 hours.

Across 36 human replies, the median wait was 235.5 minutes, about 3 hours 56 minutes. 8 came within an hour. 8 took more than a day. The quickest person replied in 4 minutes. The slowest took 3.3 days.

We sent our questions in the late morning, inside working hours. Neha shops at night. Her wait is longer.

What this means for youYou do not need to be online all night. Promise a reply within 4 hours in working time, say it on your contact page, and keep it every day.

Finding 4

Bots were fast. Not always right.

All 7 site chats we used replied, and 6 of them were bots. 8 of 10 bot replies came within 3 minutes. Bots gave 4 of the 20 full answers. Every human answer we got came by email.

But speed with wrong data hurts. 3 brands told us a product was out of stock or not listed while their own store showed it available at that moment: a coffee brand’s chat bot, a person at a snack brand, and a 27 second reply at a cosmetics brand.

What this means for youSpeed helps when the answer is true. Connect replies, from a person or a bot, to live stock, and check before anyone says “sold out”.

Finding 5

Walls went up before help.

6 brands (7.2%) asked for a phone number or an order id before they would help with a question about buying. A buyer has no order id yet. 24 brands sent a ticket mail at some point.

Along the way, 5 site chats would not let us type a question until we gave a phone number. We never share one, so those brands are not in the 83. A real buyer meets the same wall.

What this means for youLet people ask before you ask them for anything. Collect the phone number after you have helped, when they want to order.

The fix is small. Save this.

None of this needs a big team or new software. It needs one owner and a few habits.

Save this

The 4 hour reply checklist

  1. Promise a time and keep it. Answer every buying question within 4 hours in working time. Say it on your contact page.
  2. Answer every part. Stock and COD are two questions. Give two clear answers in one message.
  3. Check live stock before you say no. A wrong “sold out” sends a ready buyer to someone else.
  4. Check the pincode for them. “Check at checkout” is half an answer.
  5. Send the link. Make the next tap the buy button.
  6. Ask for nothing before you help. No phone number, no order id, no form before a buyer’s question.
  7. If you need time, say when. A holding line with a real time beats a ticket number.

Print it. Stick it next to whoever answers your inbox.

Reply scripts: copy, change the words in [brackets], send

Five replies that answer the whole question in one go. Use them in email, site chat or WhatsApp.

Stock: yes, it is available

Hi [Name], yes, the [product] in [size or shade] is in stock right now. Here is the link: [link]. COD is available to [city or pincode] too. Anything else I can check for you?

COD question

Hi [Name], yes, cash on delivery is available to [pincode]. [There is a ₹X COD fee, and prepaid orders ship free.] Here is the link to order: [link]. Happy to help if you have any other question.

Pincode and delivery

Hi [Name], I checked [pincode] for you. We deliver there in [X to Y] days, and COD is [available / not available yet] for that pincode. [If not: you can pay by UPI on the same link.] Link: [link]

Out of stock, said kindly

Hi [Name], sorry, the [product] in [size] is sold out right now. It should be back by [date], and I can message you here the day it lands. If you need it sooner, [similar product] is in stock in your size: [link].

When you need more time

Hi [Name], got your question about the [product]. I am checking stock with our team and will reply by [time today]. Thank you for waiting.

Key takeaways

  • We asked 83 Indian D2C brands one buying question in October 2026. 51 responded and 32 never replied.
  • 20 of 83 fully answered both parts. 16 answered half, most often with “check your pincode at checkout”.
  • The median human reply took about 4 hours. 8 of 36 took more than a day.
  • Bots were quick, but 3 brands said a product was unavailable while their own store showed it in stock.
  • Use the 4 hour reply checklist and the five scripts above.

Questions people ask

How fast do Indian D2C brands reply to a customer question?

In our October 2026 test of 83 brands, the median human reply took 235.5 minutes, about 4 hours. 8 of 36 human replies came within an hour and 8 took more than a day. 32 of the 83 brands never replied.

What is a good reply time for an online store in India?

Aim to answer every pre purchase question within 4 hours in working time, and sooner in the evening when people shop. In our sample, 19 of 36 human replies came within 4 hours. That is a fair bar any small team can keep.

Is 83 brands a big enough sample?

It is a small and honest sample: 83 brands, one week, one question each, mostly well known D2C names. Read it as a clear signal, not a census. The method is on this page so anyone can repeat it.

Did you ask on WhatsApp?

No. We asked through the brand's own site chat or the support email printed on its site. Routes that were WhatsApp alone, behind a login, or asked for a phone number before we could type were skipped and noted as walls.

Are chatbots good for pre purchase questions?

In our test every site chat replied, and 8 of 10 bot replies came within 3 minutes. Bots gave 4 of the 20 full answers. A bot helps when it reads live stock and your COD rules. When it does not know, it should hand over to a person.

Which brands did you test?

We do not name them. The point is the pattern, not a scoreboard. The 83 brands were fashion 18, beauty 18, food 16, jewellery 15, wellness 11 and home 5.

Can I quote these numbers?

Yes. Please cite “India Reply Report 2026, The Human Algorithm.” and link to this page. Please do not guess or name the brands.

If you want help

Start with the free check. Most people find what they need there. If you want ready replies for every stage of an enquiry, the kit is beside it.

The question is the easy part. Be the brand that answers it.

Download the full report (PDF)

How to cite: India Reply Report 2026, The Human Algorithm., https://madhurtyagi.sites.tab.bot/reply-report-2026/. Fieldwork 25 to 31 October 2026, 83 Indian D2C brands, one pre purchase question each, replies counted to 1 November 2026, 11:17 AM IST. Brands are not named. Neha is a composite. Madhur Tyagi is the founder of The Human Algorithm. 8+ years, 25+ brands, since 2017.

Keep reading