Chat does not sell for an online store on its own. It helps when it answers a question that would otherwise stop a purchase, and when that answer comes at the moment the question arises. Everything else comes down to configuration: which pages show the widget, who handles it, and what happens when no one is available on the other end.
You will not find figures here about sales increases after adding chat. Claims circulating on Polish websites — about double-digit conversion lifts or the percentage of customers expecting an answer within minutes — are repeated without identifying the study, sample, or definitions used, so they cannot be verified or applied to your store. Instead, we describe the conditions under which store chat genuinely works and the places where it does harm. Your own result from the previous quarter is the benchmark.
The prerequisite no one talks about
Guides to store chat usually start with choosing a tool. That gets the order wrong. The first decision is not about software but staffing — because a chat that looks available and goes unanswered is worse than no chat at all. A customer who writes and receives no reply leaves the store with a specific negative impression, not a neutral one.
This problem is more acute for an online store than for a service business because peak shopping traffic rarely overlaps with office hours. Evenings after dinner, Sunday afternoons, the days just after payday, and the weeks before holidays are when people browse and decide. If staffing ends at 5 p.m., chat mainly serves people who would have bought without a conversation anyway. Check your own data before implementing anything.
- At which hours and on which days of the week are your orders placed? Look at the distribution, not the daily average.
- How many of those hours does your current team cover, and how many leave no one on the other end?
- Does evening traffic have a different share of mobile users than daytime traffic? Those users ask different questions and require different support.
- Who responds on Saturdays and Sundays if weekends generate orders?
- What happens to a question asked at 10 p.m.: does it disappear, wait, or become a numbered ticket?
The final point is decisive and usually the cheapest to fix. In Meetlead, you set an availability schedule separately for each website. Outside those hours, the widget switches to offline mode with a form that creates a numbered ticket. That is more honest than promising an immediate answer no one can provide at that time. We explain how to calculate staffing needs and arrange shifts in our article on customer service team staffing.
Set up the schedule and offline mode before proactive messages and automations. Schedule implementation — we will help match chat hours to the distribution of your orders rather than your office hours.
Where to place chat and where not to show it
In an online store, the widget does not have to — and should not — behave the same way on every page. The shopping journey contains stages with different levels of sensitivity: at some, a conversation unblocks a decision; at others, it interrupts an action the customer was about to complete. The rule is simple: the closer the customer is to confirming an order, the less you should interrupt.
Scroll the table sideways to see all columns.
| Place in the store | Show chat | Why |
|---|---|---|
| Product page | Yes, prominently | This is where questions about size, availability, and delivery dates arise, and a direct answer can unblock the decision |
| Category listing | Yes, discreetly | The customer is still comparing. An icon is enough; a proactive message is distracting |
| Delivery, returns, and contact pages | Yes | Anyone who lands here has a specific question and is looking for a way to ask it |
| Cart | Carefully, without proactive messages | An icon is fine, but a pop-up invitation is not — it covers the order summary and costs |
| Order details and payment screen | No | Anything that covers a field or button can cost you a transaction that was already halfway complete |
| Order confirmation | Yes, as a support channel | Post-purchase questions will arrive anyway, so it is better to send them somewhere they will be recorded |
Implementing this table comes down to visibility rules. In Meetlead, the widget can appear on every page, only on specified path prefixes, or everywhere except specified paths. You can also select devices separately. Excluding the payment path then takes one entry in the rules rather than a change to the store's code.
Check the same layout separately on a phone — ideally your own, rather than in a desktop preview. An icon that looks discreet on a large screen can cover the buy button, price bar, or availability message on a small one.
Five questions customers really ask in online stores
Conversations in online stores are surprisingly repetitive. After two weeks of reading transcripts, you can usually see that most threads are variations on five questions. Each one says something different about the store and calls for a different response, so it is worth separating them instead of counting all conversations together. During a promotion you may also see questions about a code rejected in the cart; our guide on what to do when a discount code is not working shows how to check its rules.
- When will my order arrive?The most common question in an online store and the cheapest one to eliminate. It keeps coming back when the delivery date is vague or appears only in the cart.
- Is this in stock?A sign that availability information on the product page is unclear or untrustworthy. Meetlead does not connect to inventory; a person answers. The more often this question appears, the more expensive support becomes.
- Which size should I choose?A question about fit, not the product itself. It usually means a size chart is missing, or there is no information about how the item fits compared with a standard size.
- What if it does not fit?This is a question about risk, not procedure. The customer is checking how easily they can reverse the decision before making it.
- Can I pay another way?The customer's preferred payment method is either unavailable or not clearly shown. This is the easiest issue to confirm in order data.
The fourth question is the most powerful because it concerns a sense of security, not a product parameter. A chat answer resolves one case; a clear explanation of the terms on the page resolves every case that follows. We explain how to phrase those terms so they genuinely support a decision in our article on returns policies and purchase decisions. The returns question often comes right after the third one, about choosing a size; which measurements to share and what to ask the shopper is covered in our guide to size questions in a clothing store.
Two weeks of transcripts is enough to identify your own top five questions. Contact us, and we will show you how to tag store conversations so the list builds itself, without reviewing them manually.
Chat versus a strong product page
Conversations are expensive and do not scale. Every repeated question is a sign that the information the customer needed was not where they looked for it. Chat detects these gaps better than any analytics tool, but using it as a permanent patch for missing content is the most expensive way to run an online store.
- OBSERVATION
Twenty conversations in one month concern the delivery date of a single product
- HYPOTHESIS
The date appears only in the cart, after the customer selects a delivery method
- VERIFICATION
Session recordings show customers scrolling up and down the product page just before leaving
- DECISION
Add the date to the product page and see whether the question disappears from conversations
In practice, the distinction is simple. A conversation makes sense when the answer depends on a particular customer's situation. Content wins wherever the answer is always the same. When a shopper cannot decide which product to choose, follow the script in the guide to pre-purchase questions in chat.
Scroll the table sideways to see all columns.
| Type of question | Cheaper answer | Reason |
|---|---|---|
| Delivery time and cost | Content on the product page | The answer is identical for everyone who asks |
| Returns and complaints policy | Content plus a brief summary beside the buy button | The customer wants to know this before deciding, not after a conversation |
| Choosing a size or variant | A conversation supported by a size chart | It depends on the individual, but the chart resolves most cases |
| Comparing two models | A conversation | It requires understanding what the customer will use the product for |
| Status of an existing order | A numbered ticket | It requires investigation and a delayed response, not an immediate one |
We describe how to identify these gaps from conversations and recordings in our article on online store product pages. One rule applies: if the same question comes back for the third time in a month, it stops being a question and becomes a task for the website.
After the sale: where chat ends and a ticket begins
In an online store, post-purchase traffic can be heavier than pre-purchase traffic, and it has a completely different character. A question before purchase must be resolved immediately because the customer is on the site and may leave at any moment. A post-purchase issue — a damaged delivery, a lost parcel, a return — requires investigation and sometimes contact with the carrier, with an answer returning to the customer hours or days later. A chat window was not designed for this: the thread disappears with the browser tab, and the customer has no reference to quote.
- Identify the type of issue in the first response.A pre-purchase question stays in chat. Anything concerning an existing order moves to the next stage.
- Assign a case number before the conversation ends.The number is the one thing the customer can reference in their next contact, including by phone.
- Give a realistic deadline instead of saying ‘soon’.For cases that depend on the carrier, promise only what you control yourself.
- Keep every channel in one thread.The offline form, the conversation, and a later email should all reach the same case, not three separate ones.
- Close cases deliberately.A case closed without an answer comes back a few days later, usually in a sharper tone and as a complaint.
In Meetlead, an offline-mode form and a conversation can create a numbered ticket, so the thread does not disappear when the browser window closes. We explain when a regular email inbox stops being enough and exactly what a case number adds in our article on ticketing systems in customer service.
Chat resolves the pre-purchase question; a numbered ticket carries the post-purchase case forward. Schedule implementation — we will configure both channels so the customer never has to tell the same story twice.
If your store runs on WooCommerce, organise customer support for order questions: check the status in the store dashboard, then give the customer the appropriate next step.
Reduce repeated delivery questions with clear updates about stages and exceptions while keeping contact details visible. See how to reduce Where Is My Order tickets.
A two-week implementation plan for an online store
Launching chat in an online store does not take a quarter. It takes the right sequence: set the boundaries first, then add anything that initiates contact with the customer. The plan below assumes one person is responsible and that the only change to the store's code is embedding the widget. If the store runs on Shopify, our step-by-step Shopify chat installation guide covers adding the widget to the theme and testing it on a phone.
- Day 1: order distribution across the day and week.Identify the hours when purchases are actually made. Those hours determine the schedule, not the other way around.
- Day 2: availability schedule and offline mode.Set the website's hours and an offline form that creates a numbered ticket.
- Days 3–4: visibility rules.Show it on product and information pages, but not on the payment path. Check the mobile view separately.
- Days 5–7: support without proactive messages.For the first week, just answer. You are gathering evidence before you automate anything.
- Day 8: review questions.List recurring topics and separate those that call for a conversation from those that require better content.
- Days 9–11: improve the content.Fix the first three gaps on product pages. This is the cheapest part of the entire implementation.
- Days 12–14: one automation and measurement.Launch one proactive message with a clear condition and compare the number of identical questions with the first week.
Meetlead's automation triggers can use the page address, time spent on the page, scroll depth, and exit intent. That is enough to reach out to someone who has been hesitating on a product page for a long time without interrupting them once during payment — provided the URL rule is in place before the proactive message goes live.
- The widget does not cover the buy button or price on mobile.
- Outside support hours, the widget states clearly when the customer can expect an answer.
- No proactive message or pop-up appears along the payment path.
- Every issue concerning an existing order receives a case number.
- Recurring questions go onto a content-improvement list, not only into the conversation archive.
- After a month, you compare the number of identical questions, not just the number of conversations.
The final point is the only measure that does not lie in an online store. A rising number of conversations means nothing by itself — it may mean chat is convenient, but it may equally mean the website has stopped answering basic questions. A decline in identical questions while sales hold steady is a sign that the information has reached the right place and chat is handling the work for which it is genuinely needed.
