A customer asks "Is this in stock, and can it arrive by Friday?" and gets no answer for three hours. Most shop owners know this feeling from the other side: by the time you reply, the customer has bought elsewhere. Slow response time is one of the quietest causes of lost sales in ecommerce, because it leaves no trace. No refund, no complaint, no bad review. Just a chat window that went cold. This article looks at what research actually says about reply speed, how to estimate what it costs your own shop, and what you can do about it without hiring a night shift.
What the research says about reply speed
There is less hard data on this than marketing blogs suggest. Many "statistics" about chat response times circulate without a source. Here are the ones that hold up.
Responding within an hour matters a lot. In a 2011 Harvard Business Review study, as reported by SmartCompany, researchers sent web inquiries to 2,241 US companies and timed the replies. Only 37% answered within an hour, 23% never answered at all, and the average response time among those that did reply within 30 days was 42 hours. Firms that got back within an hour were nearly seven times as likely to qualify the lead as firms that waited even one hour longer. That study looked at sales leads in general, not online shops, but the pattern is the same one shop owners see: interest fades fast.
Answering at all lifts conversion. A study of 1,446 women's clothing sellers on Taobao, by Haoyan Sun, Jianqing Chen and Ming Fan, found that sellers who answered questions promptly in live chat had noticeably higher conversion. The sellers answered only about half of questions on average, and the researchers estimated that answering every time could raise conversion by around 9% (Lehigh Business summary).
A chat changes the purchase decision. A study by Tan, Wang and Tan in Information Systems Research used Alibaba records of browsing, live chat and purchases for Apple and Samsung tablets. It estimated that live chat raised the purchase probability for tablets by about 16% (as reported by Morung Express).
Many businesses still do not reply. When SuperOffice emailed customer service requests to 1,000 companies, 62% never replied and the average reply took 12 hours and 10 minutes. Only 20% answered the question fully in the first reply.
Customers expect "immediate". HubSpot's research reports that 90% of customers rate an "immediate" response as important or very important when they have a customer service question.
None of these numbers will predict exactly what happens in your shop. Together they say one thing clearly: speed and completeness of the answer change whether people buy.
How slow response time turns into lost sales in ecommerce
A slow reply costs you in three different ways.
1. The customer buys elsewhere
Online, the next shop is one tab away. Someone comparing three stores for the same drill will buy from the first one that confirms stock and delivery. Your price might be lower. It does not matter if nobody told them.
2. The customer gives up on the purchase
Some questions block the purchase entirely. Baymard Institute's checkout research lists the reasons US shoppers abandon a purchase: 40% because extra costs were too high, 20% because delivery was too slow, 13% because the returns policy was not satisfactory, and 12% because they could not see the total cost up front. Each of these is a question a customer could have asked in chat: "How much is delivery?", "When will it arrive?", "Can I return it?" If the answer comes tomorrow, the purchase does not happen today.
3. Your team pays in hours and mistakes
Even when you do reply, someone types "Yes, in stock, delivery 1–2 days" for the hundredth time. Repetitive questions eat the hours your managers should spend on harder conversations, and tired people quote last month's price. We look at the most common questions in detail in the typical questions customers ask an online store.
Estimate your own losses: simple math
You do not need industry averages. You can estimate the cost with your own numbers. The example below is illustrative; replace every figure with yours.
Say your shop gets 40 pre-purchase questions a day. A quarter of them, 10, arrive in the evening or at night, when nobody answers until morning. Say that with a fast answer 1 in 5 of those people would have bought, and your average order is $45.
| Step | Calculation | Result |
|---|---|---|
| Unanswered evening questions per month | 10 × 30 | 300 |
| Orders that a fast answer would have saved | 300 × 20% | 60 |
| Revenue lost per month | 60 × $45 | $2,700 |
Change the assumptions and the result moves a lot. That is the point: run it with your own chat history. To get your inputs:
- Count the questions. Export a week of chats from your widget, Telegram and email. Count only pre-purchase questions.
- Mark when they arrived. Split them into "answered within 10 minutes", "answered same day" and "answered next day".
- Check what happened. For the slow ones, did the customer order? Compare that with the share of fast-answered customers who ordered.
The gap between those two groups is your real cost of slow replies. Most shops that do this exercise are surprised by how many questions arrive after closing time. If Telegram is a big channel for you, see how to stop losing night-time inquiries.
Where the slow replies come from
Before fixing anything, find the cause. Slow replies usually come from one of four places:
- After hours. Questions at 23:00 wait until 09:00. More on this in why a 24/7 online consultant matters.
- Peaks. Black Friday, the run-up to New Year and other sales days bring more chats than any team can handle.
- Repetition. Your managers are busy with "do you have it?" while a customer with a real problem waits.
- Lookups. Each answer needs a check: open the admin panel, find the product, check stock, open the delivery page. Thirty seconds per question adds up.
What actually fixes it
Answer the repetitive layer automatically
Most pre-purchase questions have an answer that is already written somewhere: in your product feed (price, stock, specs) or on your site (delivery, payment, returns). An AI assistant that reads those sources can answer in seconds, at any hour. The important condition is that it answers only from your data. An assistant that improvises a price or a delivery term creates a new problem instead of solving one.
This is how Vendigit works. The assistant searches your product feed and your policy pages. It picks a product, and Vendigit fills in the name, price and link on the product card straight from your feed. The assistant is instructed not to state a price or stock level it didn't find. If your delivery terms are not written anywhere, it says so instead of guessing.
Hand over the rest, quickly
Some questions should go to a person: complaints, refunds, wholesale prices, a damaged delivery, an order that was already placed. The cost of a slow reply is highest here, because these customers are already unhappy. A good setup routes these to a manager at once, with the whole conversation attached. We cover the rules in when a chatbot should hand over to a human.
Keep customer data out of the loop
Speed should not come at the cost of privacy. Customers sometimes type a card number or a phone into a chat. Your assistant should never ask for that data, and should hide it from the AI model when it arrives. See AI chatbots and customer data privacy.
Measure before and after
Keep the three numbers from the exercise above: questions per day, share answered within 10 minutes, and conversion of chats into orders. Check them again a month after any change.
What this costs compared to the problem
A person answering chats in the evening is the obvious fix, and for some shops it is the right one. An assistant is usually cheaper. Vendigit's Starter plan is $20 a month for one shop and 300 dialogs on the website widget. Growth is $65 a month and adds Telegram, team members and up to 2,000 dialogs. See the plans. Compare that with the monthly loss from your own estimate, not ours. For a detailed comparison with a support agent's salary, see AI consultant cost vs a support operator.
FAQ
How fast should an online store reply to a customer?
As fast as you can, and within the hour at the latest. The HBR study found a sharp drop in successful contact after the first hour. For chat, customers expect a reply within minutes.
Do slow replies matter if my prices are the lowest?
Yes. A customer who does not get an answer does not know your price is lower, or whether the item is in stock. They buy where they got the answer.
Which questions should I automate first?
The ones with an answer in your feed or on your site: availability, price, specifications, delivery and payment terms. Complaints, refunds and wholesale requests should still go to a person.
Will an AI assistant make mistakes with prices?
Product cards get their price straight from your feed, which re-syncs every 6 hours when you add it by URL, and the assistant is instructed not to state a price it didn't find. If your feed is wrong, fix it there and the assistant follows.