Sizing questions on WhatsApp that prevent returns

Most size-related returns start with a customer who was not sure and ordered anyway. A short conversation before the order removes the guess, and WhatsApp is where that conversation can actually happen.

This is a flow that pays for itself in stock movement rather than in messages saved, which is why it is worth building even if your enquiry volume is low.

Where the flow sits

Two natural placements, and they suit different stores:

Trigger	Fits	Note
Customer asks about size	Stores where people message before ordering	Free — the conversation window is open
After an order is placed	Stores selling mostly through the website	Needs a template if it lands outside the window

The first is better and cheaper. The second is a fallback for catching an order before it ships, and it has to be worded as a check rather than a sales message.

The two questions that do most of the work

Not "what size are you". Customers frequently do not know, and asking makes them guess in writing.

1. A measurement or a reference garment. "What is your chest measurement?" or "What size do you normally take in [a well-known brand]?" Both give you something objective.
2. Fit preference. "Do you prefer a relaxed or a fitted look?" Two buttons, and it resolves the between-sizes case that produces most returns.

Two answers and you can recommend a size with a reason attached, which is the part that makes the customer confident enough to keep it.

Building it with buttons

Buttons rather than free text, because "M", "m", "medium" and "prob a medium" are four ways of saying the same thing and only one of them is usable later. WhatsApp allows three buttons, with a list when your range needs more.

Saving the answer is the whole point

Ask once, keep forever. A customer's size saved to a contact field means every future conversation starts from knowledge instead of a question, and it means your team can see it without scrolling back through an old chat.

Define the fields before you build the flow — a field that has not been defined is written somewhere you cannot see, with no error to tell you. The mechanics are in saving customer answers to contact fields.

What to do with a genuinely uncertain customer

Some people are between sizes and no question will resolve it. Three honest options, in order of preference:

• Say which way your product runs. "This style runs small, most customers take the next size up" is the single most useful sentence in the flow.
• Offer the exchange policy explicitly. Removing the risk is often what closes the order.
• Hand to a person for anything unusual. A two-minute human conversation beats a wrong parcel.

Note what is missing: encouraging them to order both sizes. It converts today and it guarantees a return, and it makes your returns rate a number you created.

Beyond clothing

The same shape works anywhere the customer has to match a product to something they own. Phone case to phone model, filter to appliance, part to vehicle, consumable to machine.

In every case the flow is the same three steps: ask what they have, save it, recommend with a reason. And in every case the saved answer makes the next purchase one question shorter.

What it costs

Nothing, when the customer started the conversation — free-form replies inside the 24-hour customer service window are free and unlimited under Meta's pricing documentation, and the window opens when they message you, per Meta's sending messages documentation.

The post-order version costs a template. Worth it if your return rate on sizing is meaningful; not worth it if you sell one-size items.

Measuring whether it worked

1. Record your return rate for size reasons for a month before you build anything.
2. Publish the flow and leave everything else alone.
3. Count the conversations that took it in the automation's activity view.
4. Compare the return rate among those customers with everyone else.
5. Read a few of the conversations — the questions people ask after your two will tell you what the third question should be.

Step five is where most of the improvement comes from, and it is free.

Frequently asked

Will customers answer two questions?

On buttons, usually. Free text is where drop-off happens.

Should the bot recommend, or a person?

The bot, for the ordinary case, with a route to a person for anything it cannot resolve. That is the balance described in bot to human handover.

Can I use AI to interpret a measurement?

To read it out of a sentence, yes. To decide the size, prefer explicit rules — and for anything that becomes a price, use the calculation block rather than a model.

What about returns for other reasons?

Different flow. The returns handling itself is in returns and exchanges on WhatsApp.

What a return actually costs you

Worth stating plainly, because it is why a two-question flow is worth the effort: a return costs you the outbound delivery, the return delivery, the handling time, the period the stock was unavailable, and frequently a customer who does not come back. The refund itself is often the smallest line.

Against that, the flow costs nothing when the customer started the conversation. There are very few interventions in ecommerce with that ratio, which is why sizing is usually the first place to look when returns are the problem.

The three questions ranked by usefulness

Question	Answer quality	Drop-off risk
"What size do you normally take in [brand]?"	High — an objective reference	Low, if buttons
"Relaxed or fitted?"	High — resolves between-sizes	Very low
"What is your chest measurement?"	Highest, when they know it	High — many will not measure

The measurement gives the best answer and loses the most people. A reasonable compromise is to offer it as an option rather than a requirement: ask the two easy questions, and let anyone who wants to be precise say so.

Two numbers worth knowing before you build

Neither is an industry statistic — they are yours, and you can get both from your own returns data in an afternoon:

1. What share of your returns cite size or fit? If it is under a tenth, this flow is not your priority.
2. Which products produce them? Returns cluster. Two or three lines usually account for most of it, and those are the ones whose descriptions should carry the "runs small" note.

A flow fixes the conversation. A corrected product description fixes the cause, and the two together work far better than either alone.

Wording that keeps people in the flow

• Ask why you are asking. "So I can make sure it fits" turns an interrogation into help.
• Give the recommendation a reason. "Large — this style runs small and you said you prefer a relaxed fit" is persuasive; "Large" is not.
• Never ask a question you will not use. If the answer changes nothing, it is friction.
• Offer the exit. Someone who wants to talk to a person should be able to say so at any point.

Those four cost nothing and account for most of the difference between a flow customers complete and one they abandon halfway.

Turning the answers into something you can use

Individual answers help one customer. The set of them helps your business, and this is the part stores rarely exploit.

After a few hundred conversations you will know which products people ask about, which sizes they hesitate between, and which of your descriptions are misleading. All of that is sitting in the contact fields the flow has been filling in.

What you can see	What to do with it
One product produces most of the questions	Fix the description or the size chart
Customers consistently size up	Say so on the product page
A size sells out and gets asked for repeatedly	A stock decision, not a messaging one
The same follow-up question keeps appearing	Add it to the flow as question three

A worked example

A store selling shirts adds two buttons. Over 4 weeks, 300 customers answer, and 2 of their 40 lines account for most of the hesitation. The description on both said "true to size"; the answers said otherwise.

Changing two product descriptions did more for the return rate than the flow did — and the flow was how they found out. That is the honest framing: this is a measurement instrument that also happens to help customers while it runs.

What not to promise

A recommendation is a recommendation. Wording it as a guarantee — "this will fit" — converts a fit problem into a complaint about being misled, and it makes the return conversation worse than if you had said nothing.

"Based on what you have told me, I would go for large" is accurate, useful, and leaves the customer's judgement intact.